2 $PostgreSQL: pgsql/doc/src/sgml/release.sgml,v 1.366 2005/09/11 23:40:33 neilc Exp $
6 [A-Z][A-Z ]+[A-Z] <command>
7 [A-Za-z_][A-Za-z0-9_]+() <function>
8 [A-Za-z_]/[A-Za-z_]+ <filename>
9 PostgreSQL <productname>
10 pg_[A-Za-z0-9_] <application>
11 [A-Z][A-Z] <type>, <envar>, <literal>
15 <appendix id="release">
16 <title>Release Notes</title>
18 <sect1 id="release-8-1">
19 <title>Release 8.1</title>
22 <title>Release date</title>
23 <simpara>2005-1?-??, Current as of 2005-08-24</simpara>
27 <title>Overview</title>
30 Major changes in this release:
37 Improve concurrent access to the shared buffer cache (Tom)
42 Access to the shared buffer cache was identified as a
43 significant scalability problem, particularly on multi-CPU
44 systems. In this release, the way that locking is done in the
45 buffer manager has been overhauled to reduce lock contention
46 and improve scalability. The buffer manager has also been
47 changed to use a <quote>clock sweep</quote> replacement
55 Allow index scans to use an intermediate in-memory bitmap (Tom)
60 In previous releases, only a single index could be used to do
61 lookups on a table. With this feature, if a query has
62 <command>WHERE tab.col1 = 4 and tab.col2 = 9</>, and there is
63 no multicolumn index on <literal>col1</> and <literal>col2</>,
64 but there is an index on <literal>col1</> and another on
65 <literal>col2</>, it is possible to search both indexes and
66 combine the results in memory, then do heap fetches for only
67 the rows matching both the <literal>col1</> and
68 <literal>col2</> restrictions. This is very useful in
69 environments that have a lot of unstructured queries where it
70 is impossible to create indexes that match all possible access
71 conditions. Bitmap scans are useful even with a single index,
72 as they reduce the amount of random access needed; a bitmap
73 index scan is efficient for retrieving fairly large fractions
74 of the complete table, whereas plain index scans are not.
81 Add two-phase commit (Heikki Linnakangas, Alvaro, Tom)
86 Two-phase commit allows transactions to be "prepared" on several
87 computers, and once all computers have successfully prepared
88 their transactions (none failed), all transactions can be
89 committed. Even if a machine crashes after a prepare, the
90 prepared transaction can be committed after the machine is
91 restarted. New syntax includes <command>PREPARE TRANSACTION</> and
92 <command>COMMIT/ROLLBACK PREPARED</>. A new system view
93 <literal>pg_prepared_xacts</> has also been added.
100 Create a new role system that replaces users and groups
106 Roles are a combination of users and groups. Like users, they
107 can have login capability, and like groups, a role can have
108 other roles as members. Roles basically remove the distinction
109 between users and groups. For example, a role can:
116 Have login capability (optionally)
128 Hold access permissions for database objects
134 Inherit permissions from other roles it is a member of
140 Once a user logs into a role, she obtains capabilities of
141 the login role plus any inherited roles, and can use
142 <command>SET ROLE</> to switch to other roles she is a member of.
143 This feature is a generalization of the SQL standard's concept of
145 This change also replaces <structname>pg_shadow</> and
146 <structname>pg_group</> by new role-capable catalogs
147 <structname>pg_authid</> and <structname>pg_auth_members</>. The old
148 tables are redefined as read-only views on the new role tables.
155 Automatically use indexes for <function>MIN()</> and
156 <function>MAX()</> (Tom)
161 In previous releases, the only way to use an index for
162 <function>MIN()</> or <function>MAX()</> was to rewrite the
163 query as <command>SELECT col FROM tab ORDER BY col LIMIT 1</>.
164 Index usage now happens automatically.
171 Move <filename>/contrib/pg_autovacuum</> into the main server
177 Integrating autovacuum into the server allows it to be
178 automatically started and stopped in sync with the database
179 server, and allows autovacuum to be configured from
180 <filename>postgresql.conf</>.
187 Add shared row level locks using <command>SELECT ... FOR SHARE</>
193 While <productname>PostgreSQL</productname>'s MVCC locking
194 allows <command>SELECT</> to never be blocked by writers and
195 therefore does not need shared row locks for typical operations,
196 shared locks are useful for applications that require shared row
197 locking. In particular this reduces the locking requirements
198 imposed by referential integrity checks.
205 Add dependencies on shared objects, specifically roles
211 This extension of the dependency mechanism prevents roles from
212 being dropped while there are still database objects they own.
213 Formerly it was possible to accidentally <quote>orphan</> objects by
214 deleting their owner. While this could be recovered from, it
215 was messy and unpleasant.
224 <title>Migration to 8.1</title>
227 A dump/restore using <application>pg_dump</application> is required
228 for those wishing to migrate data from any previous release.
232 The 8.0 release announced that the <function>to_char()</> function
233 for intervals would be removed in 8.1. However, since no better API
234 has been suggested, <function>to_char(interval)</> has been enhanced in
235 8.1 and will remain in the server.
239 Observe the following incompatibilities:
246 <varname>add_missing_from</> is now false by default (Neil)
249 By default, we now generate an error if a table is used in a query
250 without a <command>FROM</> reference. The old behavior is still
251 available, but the parameter must be set to 'true' to obtain it.
257 Cause input of a zero-length string ('') for float4/float8/oid
258 to throw an error, rather than treating it as a zero (Neil)
261 This change is consistent with the current handling of
262 zero-length strings for integers. The schedule for this change
263 was announced in 8.0.
269 <varname>default_with_oids</> is now false by default (Neil)
272 With this option set to false, user-created tables no longer
273 have an OID column unless <command>WITH OIDS</> is specified in
274 <command>CREATE TABLE</>. Though OIDs have existed in all
275 releases of <productname>PostgreSQL</>, their use is limited
276 because they are only four bytes long and the counter is shared
277 across all installed databases. The preferred way of uniquely
278 identifying rows is via sequences and the <type>SERIAL</> type,
279 which have been supported since <productname>PostgreSQL</> 6.4.
285 Add <literal>E''</> syntax so eventually ordinary strings can
286 treat backslashes literally (Bruce)
289 Currently <productname>PostgreSQL</productname> processes a
290 backslash in a string literal as introducing a special escape sequence,
291 e.g. <literal>\n</> or <literal>\010</>.
292 While this allows easy entry of special values, it is
293 non-standard and makes porting of applications from other
294 databases more difficult. For this reason, the
295 <productname>PostgreSQL</productname> project is planning to
296 remove the special meaning of backslashes in strings. For
297 backward compatibility and for users who want special backslash
298 processing, a new string syntax has been created. This new string
299 syntax is formed by writing an <literal>E</> immediately preceding the
300 single quote that starts the string, e.g. <literal>E'hi\n'</>. While
301 this release does not change the handling of backslashes in strings, it
302 does add new configuration parameters to help users migrate applications
309 <varname>standard_conforming_strings</> — does this release
310 treat backslashes literally in ordinary strings?
316 <varname>escape_string_warning</> — warn about backslashes in
317 ordinary (non-E) strings
324 The <varname>standard_conforming_strings</> value is read-only.
325 Applications can retrieve the value to know how backslashes are
326 processed. (Presence of the parameter can also be taken as an
327 indication that <literal>E''</> string syntax is supported.)
328 In a future release, <varname>standard_conforming_strings</>
329 will be true, meaning backslashes will be treated literally in
330 non-E strings. To prepare for this change, use <literal>E''</>
331 strings in places that need special backslash processing, and
332 turn on <varname>escape_string_warning</> to find additional
333 strings that need to be converted to use <literal>E''</>.
334 Also, use two single-quotes (<literal>''</>) to embed a literal
335 single-quote in a string, rather than the
336 <productname>PostgreSQL</productname>-supported syntax of
337 backslash single-quote (<literal>\'</>). The former is
338 standards-conforming and does not require the use of the
339 <literal>E''</> string syntax. You can also use the
340 <literal>$$</> string syntax, which does not treat backslashes
347 Make <command>REINDEX DATABASE</> reindex all indexes in the
351 The old behavior of <command>REINDEX</> database reindexed only
352 system tables. This new behavior seems more intuitive. A new
353 command <command>REINDEX SYSTEM</> allows for reindexing just the
360 In <application>psql</application>, treat unquoted
361 <literal>\{digit}+</> sequences as octal (Bruce)
364 In previous releases, <literal>\{digit}+</> sequences were
365 treated as decimal, and only <literal>\0{digit}+</> were treated
366 as octal. This change was made for consistency.
372 Remove grammar productions for prefix and postfix <literal>%</>
373 and <literal>^</> operators
377 These have never been documented and complicated the use of the
378 modulus operator (<literal>%</>) with negative numbers.
384 Make <literal>&<</> and <literal>&></> for polygons
385 consistent with the box "over" operators (Tom)
393 <title>Additional Changes</title>
396 Below you will find a detailed account of the additional changes
397 between <productname>PostgreSQL</productname> 8.1 and the
398 previous major release.
402 <title>Performance Improvements</title>
407 Improve GiST and rtree index performance (Neil)
413 Improve the optimizer, including auto-resizing of hash joins
420 Overhaul internal API in several areas
426 Change WAL record CRCs from 64-bit to 32-bit (Tom)
429 We determined that the extra cost of computing 64-bit CRCs was
430 significant, and the gain in reliability too marginal to justify it.
436 Prevent writing large empty gaps in WAL pages (Tom)
442 Allow non-consecutive index columns to be used in a multi-column
446 For example, this allows an index on columns a,b,c to be used in
447 a query with <command>WHERE a = 4 and c = 10</>.
453 Skip WAL logging for <command>CREATE TABLE AS</> /
454 <command>SELECT INTO</> (Simon)
457 Since a crash during <command>CREATE TABLE</> would cause the
458 table to be dropped during recovery, there is no reason to WAL
459 log as the table is loaded. (Logging still happens if WAL
460 archiving is enabled, however.)
466 Allow concurrent GIST index access (Teodor, Oleg)
472 Add configuration parameter <varname>full_page_writes</> to
473 control writing full pages to WAL (Bruce)
476 To prevent partial disk writes from corrupting the database,
477 <productname>PostgreSQL</productname> writes a complete copy of
478 each database disk page to WAL the first time it is modified
479 after a checkpoint. This option turns off that functionality for more
480 speed. This is safe to use with battery-backed disk caches where
481 partial page writes cannot happen.
487 Add <varname>constraint_exclusion</varname> configuration
488 parameter to restrict child table lookups based on table
492 This allows for a type of table partitioning. If child tables
493 store separate key ranges and this is enforced using appropriate
494 <command>CHECK</> constraints, the optimizer will skip child
495 table accesses when the constraint guarantees no matching rows
496 exist in the child table.
502 Use <literal>O_DIRECT</> if available when using
503 <literal>O_SYNC</> for <varname>wal_sync_method</varname>
507 <literal>O_DIRECT</> causes disk writes to bypass the kernel
508 cache, and for WAL writes, this improves performance.
514 Improve <command>COPY FROM</> performance (Alon Goldshuv)
517 This was accomplished by reading <command>COPY</> input in
518 larger chunks, rather than character by character.
526 <title>Server Changes</title>
531 Prevent problems due to transaction ID (XID) wraparound (Tom)
534 The server will now warn when the transaction counter approaches
535 the wraparound point. If the counter becomes too close to wraparound,
536 the server will stop accepting queries. This ensures that data is
537 not lost before needed vacuuming is performed.
543 Fix problems with object IDs (OIDs) conflicting with existing system
544 objects after the OID counter has wrapped around (Tom)
550 Add warning about the need to increase
551 <varname>max_fsm_relations</> and <varname>max_fsm_pages</>
552 during <command>VACUUM</> (Ron Mayer)
558 Add <varname>temp_buffers</> configuration parameter to allow
559 users to determine the size of the local buffer area for
560 temporary table access (Tom)
566 Add session start time and client IP address to
567 <literal>pg_stat_activity</> (Magnus)
573 Enhance <literal>pg_locks</> display (Tom)
579 Log queries for client-side <command>PREPARE</> and
580 <command>EXECUTE</> (Simon)
586 Allow Kerberos name and user name case sensitivity to be
587 specified in <filename>postgresql.conf</> (Magnus)
593 Add configuration parameter <varname>krb_server_hostname</> so
594 that the server hostname can be specified as part of service
595 principal (Todd Kover)
598 If not set, any service principal matching an entry in the
599 keytab can be used. This is new Kerberos matching behavior in
606 Add <varname>log_line_prefix</> options for millisecond
607 timestamps (<literal>%m</>) and remote host (<literal>%h</>) (Ed
614 Add WAL logging for GIST indexes (Teodor, Oleg)
617 GIST indexes are now safe for crash and point-in-time recovery.
623 Remove old <filename>*.backup</> files when we do
624 <function>pg_stop_backup()</> (Bruce)
627 This prevents a large number of <filename>*.backup</> files from
628 existing in <filename>/pg_xlog</>.
634 Add configuration parameters to control TCP/IP keep-alive
635 times for idle, interval, and count (Oliver Jowett)
639 These values can be changed to allow more rapid detection of
640 lost client connections.
646 Add per-user and per-database connection limits (Petr Jelinek)
649 Using <command>ALTER USER</> and <command>ALTER DATABASE</>,
650 limits can now be enforced on the maximum number of sessions that
651 can concurrently connect as a specific user or to a specific database.
652 Setting the limit to zero disables user or database connections.
658 Allow more than two gigabytes of shared memory and per-backend
659 work memory on 64-bit machines (Koichi Suzuki)
668 <title>Query Changes</title>
673 Add temporary views (Koju Iijima, Neil)
679 Fix <command>HAVING</> without any aggregate functions or
680 <command>GROUP BY</> so that the query returns a single group (Tom)
683 Previously, such a case would treat the <command>HAVING</>
684 clause the same as a <command>WHERE</> clause. This was not per spec.
690 Add <command>USING</> clause to allow additional tables to be
691 specified to <command>DELETE</> (Euler Taveira de Oliveira, Neil)
694 In prior releases, there was no clear method for specifying
695 additional tables to be used for joins in a <command>DELETE</>
696 statement. <command>UPDATE</> already has a <literal>FROM</>
697 clause for this purpose.
703 Add support for <literal>\x</> hex escapes in backend and ecpg
707 This is just like the standard C <literal>\x</> escape syntax.
708 Octal escapes were already supported.
714 Add <command>BETWEEN SYMMETRIC</> query syntax (Pavel Stehule)
717 This feature allows <command>BETWEEN</> comparisons without
718 requiring the first value to be less than the second. For
719 example, <command>2 BETWEEN [ASYMMETRIC] 3 AND 1</> returns
720 false, while <command>2 BETWEEN SYMMETRIC 3 AND 1</> returns
721 true. <command>BETWEEN ASYMMETRIC</> was already supported.
727 Add <command>NOWAIT</> option to <command>SELECT ... FOR
728 UPDATE/SHARE</> (Hans-Juergen Schoenig)
731 While the <varname>statement_timeout</> configuration
732 parameter allows a query taking more than a certain amount of
733 time to be cancelled, the <command>NOWAIT</> option allows a
734 query to be canceled as soon as a <command>SELECT ... FOR
735 UPDATE/SHARE</> command cannot immediately acquire a row lock.
743 <title>Object Manipulation Changes</title>
748 Track dependencies of shared objects (Alvaro)
751 <productname>PostgreSQL</productname> allows global tables
752 (users, databases, tablespaces) to reference information in
753 multiple databases. This addition adds dependency information
754 for global tables, so, for example, user ownership can be
755 tracked across databases, so a user who owns something in any
756 database can no longer be removed. Dependency tracking already
757 existed for database-local objects.
763 Allow limited <command>ALTER OWNER</> commands to be performed
764 by the object owner (Stephen Frost)
767 Prior releases allowed only superusers to change object owners.
768 Now, ownership can be transferred if the user executing the command
769 owns the object and would be able to create it as the new owner
770 (that is, the user is a member of the new owning role and that role
771 has the CREATE permission that would be needed to create the object
778 Add <command>ALTER</> object <command>SET SCHEMA</> capability
779 for some object types (tables, functions, types) (Bernd Helmle)
782 This allows objects to be moved to different schemas.
788 Add <command>ALTER TABLE ENABLE/DISABLE TRIGGER</command> to
789 disable triggers (Satoshi Nagayasu)
798 <title>Utility Command Changes</title>
803 Allow <command>TRUNCATE</> to truncate multiple tables in a
804 single command (Alvaro)
807 Because of referential integrity checks, it is not allowed to
808 truncate a table that is part of a referential integrity
809 constraint. Using this new functionality, <command>TRUNCATE</>
810 can be used to truncate such tables, if both tables involved in
811 a referential integrity constraint are truncated in a single
812 <command>TRUNCATE</> command.
818 Properly process carriage returns and line feeds in
819 <command>COPY CSV</> mode (Andrew)
822 In release 8.0, carriage returns and line feeds in <command>CSV
823 COPY TO</> were processed in an inconsistent manner. (This was
824 documented on the TODO list.)
830 Add <command>COPY WITH CSV HEADER</> to allow a header line as
831 the first line in <command>COPY</> (Andrew)
834 This allows handling of the common <command>CSV</> usage of
835 placing the column names on the first line of the data file. For
836 <command>COPY TO</>, the first line contains the column names,
837 and for <command>COPY FROM</>, the first line is ignored.
843 On Windows, display better sub-second precision in
844 <command>EXPLAIN ANALYZE</> (Magnus)
850 Add trigger duration display to <command>EXPLAIN ANALYZE</>
854 Prior releases included trigger execution time as part of the
855 total execution time, but did not show it separately. It is now
856 possible to see how much time is spent in each trigger.
862 Add support for <literal>\x</> hex escapes in <command>COPY</>
866 Previous releases only supported octal escapes.
872 Make <command>SHOW ALL</> include variable descriptions
876 <command>SHOW</> varname still only displays the variable's
877 value and does not include the description.
883 Make <application>initdb</application> create a new standard
884 database called <literal>postgres</>, and convert utilities to
885 use <literal>postgres</> rather than <literal>template1</> for
886 standard lookups (Dave)
889 In prior releases, <literal>template1</> was used both as a
890 default connection for utilities like
891 <application>createuser</application>, and as a template for
892 new databases. This caused <command>CREATE DATABASE</> to
893 sometimes fail, because a new database cannot be created if
894 anyone else is in the template database. With this change, the
895 default connection database is now <literal>postgres</>,
896 meaning it is much less likely someone will be using
897 <literal>template1</> during <command>CREATE DATABASE</>.
903 Create new <application>reindexdb</application> command-line
904 utility by moving <filename>/contrib/reindexdb</> into the
905 server (Euler Taveira de Oliveira)
914 <title>Data Type and Function Changes</title>
919 Add <function>MAX()</> and <function>MIN()</> aggregates for
920 array types (Koju Iijima)
926 Fix <function>to_date()</> and <function>to_timestamp()</> to
927 behave reasonably when <literal>CC</> and <literal>YY</> fields
928 are both used (Karel Zak)
931 If the format specification contains <literal>CC</> and a year
932 specification is <literal>YYY</> or longer, ignore the
933 <literal>CC</>. If the year specification is <literal>YY</> or
934 shorter, interpret <literal>CC</> as the previous century.
940 Add <function>md5(bytea)</> (Abhijit Menon-Sen)
943 <function>md5(text)</> already existed.
949 Fix <type>CHAR()</> to properly pad spaces to the specified
950 length when using a multiple-byte character set (Yoshiyuki
954 In prior releases, the padding of <type>CHAR()</> was incorrect
955 because it only padded to the specified number of bytes without
956 considering how many characters were stored.
962 Add support for <command>numeric ^ numeric</> based on
963 <function>power(numeric, numeric)</>
966 The function already existed, but there was no operator assigned
973 Fix <type>NUMERIC</> modulus by properly truncating the quotient
974 during computation (Bruce)
977 In previous releases, modulus for large values sometimes
978 returned negative results due to rounding of the quotient.
984 Add a function <function>lastval()</>(Dennis Björklund)
987 <function>lastval()</> is a simplified version of
988 <function>currval()</>. It automatically determines the proper
989 sequence name based on the most recent <function>nextval()</> or
990 <function>setval()</> call performed by the current session.
996 Add <function>to_timestamp(DOUBLE PRECISION)</>
999 Converts Unix seconds since 1970 to a <type>TIMESTAMP WITH
1006 Add <function>pg_postmaster_start_time()</> function (Euler
1007 Taveira de Oliveira, Matthias Schmidt)
1013 Allow the full use of time zone names in <command>AT TIME
1014 ZONE</>, not just the short list previously available (Magnus)
1017 Previously, only a predefined list of time zone names were
1018 supported by <command>AT TIME ZONE</>. Now any supported time
1019 zone name can be used, e.g.:
1021 SELECT CURRENT_TIMESTAMP AT TIME ZONE 'Europe/London';
1023 In the above query, the time zone used is adjusted based on the
1024 daylight savings time rules that were in effect on the supplied
1031 Add <function>GREATEST()</> and <function>LEAST()</> variadic
1032 functions (Pavel Stehule)
1035 These functions take a variable number of arguments and return
1036 the greatest or least value among the arguments.
1042 Add <function>pg_column_size()</> (Mark Kirkwood)
1045 This returns storage size of a column, including any
1052 Add <function>regexp_replace()</> (Atsushi Ogawa)
1055 This allows regular expression replacement, like sed. An optional
1056 flag argument allows selection of global (replace all) and
1057 case-insensitive modes.
1063 Fix interval division and multiplication (Bruce)
1066 Previous versions sometimes returned unjustified results, like
1067 <command>'4 months'::interval / 5</> returning <command>'1 mon
1074 Add an internal day field to <type>INTERVAL</> so a one day
1075 interval can be distinguished from a 24 hour interval (Michael
1079 Days that contain a daylight savings time adjustment are not 24
1080 hours, but typically 23 or 25 hours. This change allows days
1081 (not fixed 24-hour periods) to be added to dates who's result
1082 includes a daylight savings time adjustment period. Therefore,
1083 while in previous releases <literal>1 day</> and <literal>24
1084 hours</> were interchangeable interval values, in this release
1085 they are treated differently, e.g.
1087 '2005-05-03 00:00:00 EST' + '1 day' = '2005-05-04 00:00:00-04'
1088 '2005-05-03 00:00:00 EST' + '24 hours' = '2005-05-04 01:00:00-04'
1095 Add <function>justify_days()</> and <function>justify_hours()</>
1096 (Michael Glaesemann)
1099 These functions, respectively, adjust days to an appropriate
1100 number of full months and days, and adjust hours to an
1101 appropriate number of full days and hours.
1107 Move <filename>/contrib/dbsize</> into the backend, and rename
1108 some of the functions (Dave Page, Andreas Pflug)
1115 <function>pg_tablespace_size()</>
1121 <function>pg_database_size()</>
1127 <function>pg_relation_size()</>
1133 <function>pg_complete_relation_size()</>
1139 <function>pg_size_pretty()</>
1146 <function>pg_complete_relation_size()</> includes indexes and TOAST
1153 Add functions for read-only file access to the cluster directory
1154 (Dave Page, Andreas Pflug)
1161 <function>pg_stat_file()</>
1167 <function>pg_read_file()</>
1173 <function>pg_ls_dir()</>
1183 Add <function>pg_reload_conf()</> to force reloading of the
1184 configuration files (Dave Page, Andreas Pflug)
1190 Add <function>pg_rotate_logfile()</> to force rotation of the
1191 server log file (Dave Page, Andreas Pflug)
1197 Change <literal>pg_stat_*</> views to include TOAST tables (Tom)
1206 <title>Encoding and Locale Changes</title>
1211 Rename some encodings to be more consistent and to follow
1212 international standards (Bruce)
1219 <literal>UNICODE</> is now <literal>UTF8</>
1225 <literal>ALT</> is now <literal>WIN866</>
1231 <literal>WIN</> is now <literal>WIN1251</>
1237 <literal>TCVN</> is now <literal>WIN1258</>
1245 The original names still work.
1251 Add support for <literal>WIN1252</> encoding (Roland Volkmann)
1257 Add support for 3 and 4-byte <literal>UTF8</> characters (John
1261 Previously only one and two-byte <literal>UTF8</> characters
1262 were supported. This is particularly important for support for
1263 some Chinese characters.
1269 Allow direct conversion between <literal>EUC_JP</> and
1270 <literal>SJIS</> to improve performance (Atsushi Ogawa)
1276 Allow the UTF8 encoding to work on Windows (Magnus)
1279 This is done by mapping UTF8 to the Windows-native UTF16
1289 <title>General Server-Side Language Changes</title>
1294 Fix <command>ALTER LANGUAGE RENAME</> (Sergey Yatskevich)
1300 Allow function characteristics, like strictness and volatility,
1301 to be modified via <command>ALTER FUNCTION</> (Neil)
1307 Increase the maximum number of function arguments to 100 (Tom)
1313 Allow SQL and PL/PgSQL functions to use <command>OUT</> and
1314 <command>INOUT</> parameters (Tom)
1317 <command>OUT</> is an alternate way for a function to return
1318 values. Instead of using <command>RETURN</>, values can be
1319 returned by assigning to parameters declared as <command>OUT</> or
1320 <command>INOUT</>. This is notationally simpler in some cases,
1321 particularly so when multiple values need to be returned.
1322 While returning multiple values from a function
1323 was possible in previous releases, this greatly simplifies the
1324 process. (The feature will be extended to other server-side
1325 languages in future releases.)
1331 Move language handler functions into the <literal>pg_catalog</> schema
1334 This makes it easier to drop the public schema if desired.
1342 <title>PL/PgSQL Server-Side Language Changes</title>
1347 Reduce memory usage of PL/PgSQL functions (Neil)
1350 The parsetree of each function is now stored in a separate
1351 memory context. This allows the memory to be easily reclaimed
1352 when it is no longer needed.
1358 Check function syntax at <command>CREATE FUNCTION</> time,
1359 rather than at runtime (Neil)
1362 Previously, most syntax errors were reported only when the
1363 function was executed.
1369 Allow <command>OPEN</> to open non-<command>SELECT</> queries
1370 like <command>EXPLAIN</> and <command>SHOW</> (Tom)
1376 No longer require functions to issue a <command>RETURN</>
1380 This is a byproduct of the newly added <command>OUT</> and
1381 <command>INOUT</> functionality. <command>RETURN</> can
1382 be omitted when it is not needed to provide the function result.
1388 Add support for an optional <command>INTO</> clause to
1389 PL/PgSQL's <command>EXECUTE</> command (Pavel Stehule, Neil)
1395 Make <command>CREATE TABLE AS</> set <command>ROW_COUNT</> (Tom)
1401 Define <literal>SQLSTATE</> and <literal>SQLERRM</> to return
1402 the <literal>SQLSTATE</> and error message of the current
1403 exception (Pavel Stehule, Neil)
1406 These variables are only accessible inside exception blocks.
1412 Allow the parameters to the <command>RAISE</> statement to be
1413 expressions (Pavel Stehule, Neil)
1419 Add a loop <command>CONTINUE</> statement (Pavel Stehule, Neil)
1425 Allow block and loop labels (Pavel Stehule)
1434 <title>PL/Perl Server-Side Language Changes</title>
1439 Allow large result sets to be returned efficiently (Abhijit
1443 This allows functions to use <function>return_next()</> to avoid
1444 building the entire result set in memory.
1450 Allow one-row-at-a-time retrieval of query results (Abhijit Menon-Sen)
1453 This allows functions to use <function>spi_query()</> and
1454 <function>spi_fetchrow()</> to avoid accumulating the entire
1455 result set in memory.
1461 Force PL/Perl to handle strings as <literal>UTF8</> if the
1462 server encoding is <literal>UTF8</> (David Kamholz)
1468 Add a validator function for PL/Perl (Andrew)
1471 This allows syntax errors to be reported at definition time,
1472 rather than execution time.
1478 Allow PL/Perl to return a Perl array when the function returns
1479 an array type (Andrew)
1482 This basically maps <productname>PostgreSQL</productname> arrays
1489 Allow Perl non-fatal warnings to generate <command>NOTICE</>
1496 Allow Perl's <literal>strict</> mode to be enabled (Andrew)
1505 <title><application>psql</> Changes</title>
1510 Add <command>\set ON_ERROR_ROLLBACK</> to allow statements in
1511 a transaction to error without affecting the rest of the
1512 transaction (Greg Sabino Mullane)
1515 This is basically implemented by wrapping every statement in a
1522 Add support for <literal>\x</> hex strings in
1523 <application>psql</> variables (Bruce)
1526 Octal escapes were already supported.
1532 Add support for <command>troff -ms</> output format (Roger
1539 Allow the history file location to be controlled by
1540 <envar>HISTFILE</> (Andreas Seltenreich)
1543 This allows configuration of per-database history storage.
1549 Prevent <command>\x</> (expanded mode) from affecting
1550 backslash-command displays (Neil)
1556 Add <option>-L</> option to <application>psql</application> to
1557 log sessions (Lorne Sunley)
1560 This option was added because some operating systems do not have
1561 simple command-line activity logging functionality.
1567 Make <command>\d</> show the tablespaces of indexes (Qingqing
1574 Allow <application>psql</application> help (<command>\h</>) to
1575 make a best guess on the proper help information (Greg Sabino
1579 This allows the user to just add <command>\h</> to the front of
1580 the syntax error query and get help on the supported syntax.
1581 Previously any additional query text beyond the command name
1582 had to be removed to use <command>\h</>.
1588 Add <command>\pset numericlocale</> to allow numbers to be
1589 output in a locale-aware format (Eugen Nedelcu)
1592 For example, using <literal>C</> locale <literal>100000</> would
1593 be output as <literal>100,000.0</> while a European locale might
1594 output this value as <literal>100.000,0</>.
1603 <title><application>pg_dump</> Changes</title>
1608 Add <option>-n</> / <option>--schema</> switch to
1609 <application>pg_restore</> (Richard van den Berg)
1612 This allows only objects from a specified schema to be restored.
1618 Allow <application>pg_dump</> to dump a consistent snapshot of
1625 Allow <application>pg_dump</> to dump large objects even in
1632 Dump comments for large objects (Tom)
1638 Add <option>--encoding</> to <application>pg_dump</>
1642 This allows a database to be dumped in an encoding that is
1643 different from the server's encoding. This is valuable when
1644 transferring the dump to a machine with a different encoding.
1653 <title><application>libpq</application> Changes</title>
1658 Add a <envar>PGPASSFILE</> environment variable to specify the
1659 password file's filename (Andrew)
1665 Add <function>lo_create()</>, that is similar to
1666 <function>lo_creat()</> but allows the OID of the large object
1667 to be specified (Tom)
1676 <title>Source Code Changes</title>
1681 Add spinlock support for the Itanium processor using Intel
1682 compiler (Vikram Kalsi)
1688 Add Kerberos 5 support for Windows (Magnus)
1694 Add Chinese FAQ (laser@pgsqldb.com)
1700 Rename Rendezvous to Bonjour to match OS/X feature renaming
1707 Add support for <literal>fsync_writethrough</literal> on
1708 Darwin (Chris Campbell)
1714 Streamline the passing of information within the server, the
1715 optimizer, and the lock system (Tom)
1721 Allow <application>pg_config</> to be compiled using MSVC (Andrew)
1724 This is required to build DBD::Pg using <application>MSVC</>.
1730 Remove support for Kerberos V4 (Magnus)
1733 Kerberos 4 had security vulnerabilities and is no longer being
1740 Code cleanups (Coverity static analysis performed by
1747 Modify <filename>postgresql.conf</> to use documentation defaults
1748 <literal>on</>/<literal>off</> rather than
1749 <literal>true</>/<literal>false</> (Bruce)
1755 Enhance <application>pg_config</> to be able to report more
1756 build-time values (Tom)
1762 Allow <application>libpq</application> to be built thread-safe
1763 on Windows (Dave Page)
1769 Allow IPv6 connections to be used on Win32 (Andrew)
1778 <title>Contrib Changes</title>
1783 Add <filename>/contrib/pg_buffercache</> contrib module (Mark
1787 This displays the contents of the buffer cache, for debugging and
1788 performance tuning purposes.
1794 Remove <filename>/contrib/array</> because it is obsolete (Tom)
1800 Clean up the <filename>/contrib/lo</> module (Tom)
1806 Move <filename>/contrib/findoidjoins</> to
1807 <filename>/src/tools</> (Tom)
1813 Remove the <literal><<</>, <literal>>></>,
1814 <literal>&<</>, and <literal>&></> operators from
1815 <filename>/contrib/cube</>
1818 These operators were not useful.
1824 Improve <filename>/contrib/btree_gist</> (Janko Richter)
1830 Improve <filename>/contrib/pgcrypto</> (Marko Kreen)
1837 Implementation of OpenPGP symmetric-key and public-key encryption
1840 Both RSA and Elgamal public-key algorithms are supported.
1846 Stand alone build: include SHA256/384/512 hashes, Fortuna PRNG
1852 OpenSSL build: support 3DES, use internal AES with OpenSSL < 0.9.7
1858 Take build parameters (OpenSSL, zlib) from <filename>configure</> result
1861 No need to edit the <filename>Makefile</> anymore.
1867 Remove support for <filename>libmhash</> and <filename>libmcrypt</>
1880 <sect1 id="release-8-0-3">
1881 <title>Release 8.0.3</title>
1884 <title>Release date</title>
1885 <simpara>2005-05-09</simpara>
1889 This release contains a variety of fixes from 8.0.2, including several
1890 security-related issues.
1894 <title>Migration to version 8.0.3</title>
1897 A dump/restore is not required for those running 8.0.X. However,
1898 it is one possible way of handling two significant security problems
1899 that have been found in the initial contents of 8.0.X system
1900 catalogs. A dump/initdb/reload sequence using 8.0.3's initdb will
1901 automatically correct these problems.
1905 The larger security problem is that the built-in character set encoding
1906 conversion functions can be invoked from SQL commands by unprivileged
1907 users, but the functions were not designed for such use and are not
1908 secure against malicious choices of arguments. The fix involves changing
1909 the declared parameter list of these functions so that they can no longer
1910 be invoked from SQL commands. (This does not affect their normal use
1911 by the encoding conversion machinery.)
1915 The lesser problem is that the <filename>contrib/tsearch2</> module
1916 creates several functions that are improperly declared to return
1917 <type>internal</> when they do not accept <type>internal</> arguments.
1918 This breaks type safety for all functions using <type>internal</>
1923 It is strongly recommended that all installations repair these errors,
1924 either by initdb or by following the manual repair procedure given
1925 below. The errors at least allow unprivileged database users to crash
1926 their server process, and may allow unprivileged users to gain the
1927 privileges of a database superuser.
1931 If you wish not to do an initdb, perform the same manual repair
1932 procedures shown in the <link linkend="release-7-4-8">7.4.8 release
1938 <title>Changes</title>
1941 <listitem><para>Change encoding function signature to prevent
1942 misuse</para></listitem>
1943 <listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
1944 <type>INTERNAL</> function results</para></listitem>
1945 <listitem><para>Guard against incorrect second parameter to
1946 <function>record_out</></para></listitem>
1947 <listitem><para>Repair ancient race condition that allowed a transaction to be
1948 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
1949 than for other purposes</para>
1950 <para>This is an extremely serious bug since it could lead to apparent
1951 data inconsistencies being briefly visible to applications.</para></listitem>
1952 <listitem><para>Repair race condition between relation extension and
1954 <para>This could theoretically have caused loss of a page's worth of
1955 freshly-inserted data, although the scenario seems of very low probability.
1956 There are no known cases of it having caused more than an Assert failure.
1958 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
1960 The comparison code was wrong in the case where the
1961 <literal>--enable-integer-datetimes</> configuration switch had been used.
1962 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
1963 it will need to be <command>REINDEX</>ed after installing this update, because
1964 the fix corrects the sort order of column values.
1966 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
1967 <type>TIME WITH TIME ZONE</> values</para></listitem>
1968 <listitem><para>Fix mis-display of negative fractional seconds in
1969 <type>INTERVAL</> values</para>
1971 This error only occurred when the
1972 <literal>--enable-integer-datetimes</> configuration switch had been used.
1974 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
1975 correctly (Neil)</para></listitem>
1976 <listitem><para>Still more 64-bit fixes for
1977 <filename>contrib/intagg</></para></listitem>
1978 <listitem><para>Prevent incorrect optimization of functions returning
1979 <type>RECORD</></para></listitem>
1980 <listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
1981 <listitem><para>Fix Borland makefile for libpq</para></listitem>
1982 <listitem><para>Fix <filename>contrib/btree_gist</> for <type>timetz</> type
1983 (Teodor)</para></listitem>
1984 <listitem><para>Make <command>pg_ctl</> check the PID found in
1985 <filename>postmaster.pid</> to see if it is still a live
1986 process</para></listitem>
1987 <listitem><para>Fix <command>pg_dump</>/<command>pg_restore</> problems caused
1988 by addition of dump timestamps</para></listitem>
1989 <listitem><para>Fix interaction between materializing holdable cursors and
1990 firing deferred triggers during transaction commit</para></listitem>
1991 <listitem><para>Fix memory leak in SQL functions returning pass-by-reference
1992 data types</para></listitem>
1998 <sect1 id="release-8-0-2">
1999 <title>Release 8.0.2</title>
2002 <title>Release date</title>
2003 <simpara>2005-04-07</simpara>
2007 This release contains a variety of fixes from 8.0.1.
2011 <title>Migration to version 8.0.2</title>
2014 A dump/restore is not required for those running 8.0.*.
2015 This release updates the major version number of the
2016 <productname>PostgreSQL</productname> libraries, so it might be
2017 necessary to re-link some user applications if they cannot
2018 find the properly-numbered shared library.
2023 <title>Changes</title>
2026 <listitem><para>Increment the major version number of all interface
2027 libraries (Bruce)</para>
2029 This should have been done in 8.0.0. It is required so 7.4.X versions
2030 of PostgreSQL client applications, like <application>psql</>,
2031 can be used on the same machine as 8.0.X applications. This might require
2032 re-linking user applications that use these libraries.
2034 <listitem><para>Add Windows-only <varname>wal_sync_method</> setting of
2035 <option>fsync_writethrough</> (Magnus, Bruce)</para>
2037 This setting causes <productname>PostgreSQL</productname> to write through
2038 any disk-drive write cache when writing to WAL.
2039 This behavior was formerly called <option>fsync</>, but was
2040 renamed because it acts quite differently from <option>fsync</> on other
2044 <listitem><para>Enable the <varname>wal_sync_method</> setting of
2045 <option>open_datasync</> on Windows, and make it the default for that
2046 platform (Magnus, Bruce)</para>
2048 Because the default is no longer <option>fsync_writethrough</>,
2049 data loss is possible during a power failure if the disk drive has
2050 write caching enabled. To turn off the write cache on Windows,
2051 from the <application>Device Manager</>, choose the drive properties,
2052 then <literal>Policies</>.
2055 <listitem><para>New cache management algorithm <acronym>2Q</> replaces
2056 <acronym>ARC</> (Tom)</para>
2058 This was done to avoid a pending US patent on <acronym>ARC</>. The
2059 <acronym>2Q</> code might be a few percentage points slower than
2060 <acronym>ARC</> for some work loads. A better cache management algorithm
2063 <listitem><para>Planner adjustments to improve behavior on freshly-created
2064 tables (Tom)</para></listitem>
2065 <listitem><para>Allow plpgsql to assign to an element of an array that is
2066 initially <literal>NULL</> (Tom)</para>
2068 Formerly the array would remain <literal>NULL</>, but now it becomes a
2069 single-element array. The main SQL engine was changed to handle
2070 <command>UPDATE</> of a null array value this way in 8.0, but the similar
2071 case in plpgsql was overlooked.
2074 <listitem><para>Convert <literal>\r\n</> and <literal>\r</> to <literal>\n</>
2075 in plpython function bodies (Michael Fuhr)</para>
2077 This prevents syntax errors when plpython code is written on a Windows or
2081 <listitem><para>Allow SPI cursors to handle utility commands that return rows,
2082 such as <command>EXPLAIN</> (Tom)</para></listitem>
2083 <listitem><para>Fix <command>CLUSTER</> failure after <command>ALTER TABLE
2084 SET WITHOUT OIDS</> (Tom)</para></listitem>
2085 <listitem><para>Reduce memory usage of <command>ALTER TABLE ADD COLUMN</>
2086 (Neil)</para></listitem>
2087 <listitem><para>Fix <command>ALTER LANGUAGE RENAME</> (Tom)</para></listitem>
2088 <listitem><para>Document the Windows-only <literal>register</> and
2089 <literal>unregister</> options of <application>pg_ctl</> (Magnus)</para></listitem>
2090 <listitem><para>Ensure operations done during backend shutdown are counted by
2091 statistics collector</para>
2093 This is expected to resolve reports of <application>pg_autovacuum</>
2094 not vacuuming the system catalogs often enough — it was not being
2095 told about catalog deletions caused by temporary table removal during
2098 <listitem><para>Change the Windows default for configuration parameter
2099 <varname>log_destination</> to <option>eventlog</> (Magnus)</para>
2101 By default, a server running on Windows will now send log output to the
2102 Windows event logger rather than standard error.
2104 <listitem><para>Make Kerberos authentication work on Windows (Magnus)</para></listitem>
2105 <listitem><para>Allow <command>ALTER DATABASE RENAME</> by superusers
2106 who aren't flagged as having CREATEDB privilege (Tom)</para></listitem>
2107 <listitem><para>Modify WAL log entries for <command>CREATE</> and
2108 <command>DROP DATABASE</> to not specify absolute paths (Tom)</para>
2109 <para>This allows point-in-time recovery on a different machine with possibly
2110 different database location. Note that <command>CREATE TABLESPACE</> still
2111 poses a hazard in such situations.
2113 <listitem><para>Fix crash from a backend exiting with an open transaction
2114 that created a table and opened a cursor on it (Tom)</para></listitem>
2115 <listitem><para>Fix <function>array_map()</> so it can call PL functions
2116 (Tom)</para></listitem>
2117 <listitem><para>Several <filename>contrib/tsearch2</> and
2118 <filename>contrib/btree_gist</> fixes (Teodor)
2120 <listitem><para>Fix crash of some <filename>contrib/pgcrypto</>
2121 functions on some platforms (Marko Kreen)</para></listitem>
2122 <listitem><para>Fix <filename>contrib/intagg</> for 64-bit platforms
2123 (Tom)</para></listitem>
2124 <listitem><para>Fix ecpg bugs in parsing of <command>CREATE</> statement
2125 (Michael)</para></listitem>
2126 <listitem><para>Work around gcc bug on powerpc and amd64 causing problems in
2127 ecpg (Christof Petig)</para></listitem>
2128 <listitem><para>Do not use locale-aware versions of <function>upper()</>,
2129 <function>lower()</>, and <function>initcap()</> when the locale is
2130 <literal>C</> (Bruce)</para>
2132 This allows these functions to work on platforms that generate errors
2133 for non-7-bit data when the locale is <literal>C</>.
2135 <listitem><para>Fix <function>quote_ident()</> to quote names that match keywords (Tom)</para></listitem>
2136 <listitem><para>Fix <function>to_date()</> to behave reasonably when
2137 <literal>CC</> and <literal>YY</> fields are both used (Karel)</para></listitem>
2138 <listitem><para>Prevent <function>to_char(interval)</> from failing
2139 when given a zero-month interval (Tom)</para></listitem>
2140 <listitem><para>Fix wrong week returned by <function>date_trunc('week')</>
2143 <function>date_trunc('week')</>
2144 returned the wrong year for the first few days of January in some years.
2146 <listitem><para>Use the correct default mask length for class <literal>D</>
2147 addresses in <type>INET</> data types (Tom)</para></listitem>
2153 <sect1 id="release-8-0-1">
2154 <title>Release 8.0.1</title>
2157 <title>Release date</title>
2158 <simpara>2005-01-31</simpara>
2162 This release contains a variety of fixes from 8.0.0, including several
2163 security-related issues.
2167 <title>Migration to version 8.0.1</title>
2170 A dump/restore is not required for those running 8.0.0.
2175 <title>Changes</title>
2178 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
2180 On platforms that will automatically execute initialization functions of a
2181 shared library (this includes at least Windows and ELF-based Unixen),
2182 <command>LOAD</> can be used to make the server execute arbitrary code.
2183 Thanks to NGS Software for reporting this.</para></listitem>
2184 <listitem><para>Check that creator of an aggregate function has the right to
2185 execute the specified transition functions</para>
2187 This oversight made it possible to bypass denial of EXECUTE
2188 permission on a function.</para></listitem>
2189 <listitem><para>Fix security and 64-bit issues in
2190 contrib/intagg</para></listitem>
2191 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
2192 Jurka)</para></listitem>
2193 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
2194 many parameters (Neil)</para></listitem>
2195 <listitem><para>Make <command>ALTER TABLE ADD COLUMN</> enforce domain
2196 constraints in all cases</para></listitem>
2197 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
2199 The result of the join was mistakenly supposed to be sorted the same as the
2200 left input. This could not only deliver mis-sorted output to the user, but
2201 in case of nested merge joins could give outright wrong answers.
2203 <listitem><para>Improve planning of grouped aggregate queries</para></listitem>
2204 <listitem><para><command>ROLLBACK TO <replaceable>savepoint</></command>
2205 closes cursors created since the savepoint</para></listitem>
2206 <listitem><para>Fix inadequate backend stack size on Windows</para></listitem>
2207 <listitem><para>Avoid SHGetSpecialFolderPath() on Windows
2208 (Magnus)</para></listitem>
2209 <listitem><para>Fix some problems in running pg_autovacuum as a Windows
2210 service (Dave Page)</para></listitem>
2211 <listitem><para>Multiple minor bug fixes in
2212 pg_dump/pg_restore</para></listitem>
2213 <listitem><para>Fix ecpg segfault with named structs used in
2214 typedefs (Michael)</para></listitem>
2220 <sect1 id="release-8-0">
2221 <title>Release 8.0</title>
2224 <title>Release date</title>
2225 <simpara>2005-01-19</simpara>
2229 <title>Overview</title>
2232 Major changes in this release:
2238 Microsoft Windows Native Server
2243 This is the first <productname>PostgreSQL</productname> release
2244 to run natively on <trademark class=registered>Microsoft Windows</> as
2245 a server. It can run as a <productname>Windows</> service. This
2246 release supports NT-based Windows releases like
2247 <productname>Windows 2000</>, <productname>Windows XP</>, and
2248 <productname>Windows 2003</>. Older releases like
2249 <productname>Windows 95</>, <productname>Windows 98</>, and
2250 <productname>Windows ME</> are not supported because these operating
2251 systems do not have the infrastructure to support
2252 <productname>PostgreSQL</productname>. A separate installer
2253 project has been created to ease installation on
2254 <productname>Windows</> — see <ulink
2255 url="http://www.postgresql.org/ftp/win32/"></ulink>.
2259 Although tested throughout our release cycle, the Windows port
2260 does not have the benefit of years of use in production
2261 environments that <productname>PostgreSQL</productname> has on
2262 Unix platforms. Therefore it should be treated with the same
2263 level of caution as you would a new product.
2267 Previous releases required the Unix emulation toolkit
2268 <productname>Cygwin</> in order to run the server on Windows
2269 operating systems. <productname>PostgreSQL</productname> has
2270 supported native clients on Windows for many years.
2282 Savepoints allow specific parts of a transaction to be aborted
2283 without affecting the remainder of the transaction. Prior
2284 releases had no such capability; there was no way to recover
2285 from a statement failure within a transaction except by
2286 aborting the whole transaction. This feature is valuable for
2287 application writers who require error recovery within a
2288 complex transaction.
2295 Point-In-Time Recovery
2300 In previous releases there was no way to recover from disk
2301 drive failure except to restore from a previous backup or use
2302 a standby replication server. Point-in-time recovery allows
2303 continuous backup of the server. You can recover either to
2304 the point of failure or to some transaction in the past.
2316 Tablespaces allow administrators to select different file systems
2317 for storage of individual tables, indexes, and databases.
2318 This improves performance and control over disk space
2319 usage. Prior releases used <application>initlocation</> and
2320 manual symlink management for such tasks.
2327 Improved Buffer Management, <command>CHECKPOINT</command>,
2328 <command>VACUUM</command>
2333 This release has a more intelligent buffer replacement strategy,
2334 which will make better use of available shared buffers and
2335 improve performance. The performance impact of vacuum and
2336 checkpoints is also lessened.
2348 A column's data type can now be changed with <command>ALTER
2356 New Perl Server-Side Language
2361 A new version of the <application>plperl</> server-side language now
2362 supports a persistent shared storage area, triggers, returning records
2363 and arrays of records, and SPI calls to access the database.
2370 Comma-separated-value (CSV) support in <command>COPY</command>
2375 <command>COPY</command> can now read and write
2376 comma-separated-value files. It has the flexibility to
2377 interpret non-standard quoting and separation characters too.
2386 <title>Migration to version 8.0</title>
2389 A dump/restore using <application>pg_dump</application> is
2390 required for those wishing to migrate data from any previous
2395 Observe the following incompatibilities:
2402 In <option>READ COMMITTED</> serialization mode, volatile functions
2403 now see the results of concurrent transactions committed up to the
2404 beginning of each statement within the function, rather than up to the
2405 beginning of the interactive command that called the function.
2411 Functions declared <option>STABLE</> or <option>IMMUTABLE</> always
2412 use the snapshot of the calling query, and therefore do not see the
2413 effects of actions taken after the calling query starts, whether in
2414 their own transaction or other transactions. Such a function must be
2415 read-only, too, meaning that it cannot use any SQL commands other than
2422 Non-deferred <option>AFTER</> triggers are now fired immediately
2423 after completion of the triggering query, rather than upon
2424 finishing the current interactive command. This makes a
2425 difference when the triggering query occurred within a function:
2426 the trigger is invoked before the function proceeds to its next
2433 Server configuration parameters <varname>virtual_host</> and
2434 <varname>tcpip_socket</> have been replaced with a more general
2435 parameter <varname>listen_addresses</>. Also, the server now listens on
2436 <literal>localhost</> by default, which eliminates the need for the
2437 <literal>-i</> postmaster switch in many scenarios.
2443 Server configuration parameters <varname>SortMem</> and
2444 <varname>VacuumMem</> have been renamed to <varname>work_mem</>
2445 and <varname>maintenance_work_mem</> to better reflect their
2446 use. The original names are still supported in
2447 <command>SET</command> and <command>SHOW</command>.
2453 Server configuration parameters <varname>log_pid</>,
2454 <varname>log_timestamp</>, and <varname>log_source_port</> have been
2455 replaced with a more general parameter <varname>log_line_prefix</>.
2461 Server configuration parameter <varname>syslog</> has been
2462 replaced with a more logical <varname>log_destination</> variable to
2463 control the log output destination.
2469 Server configuration parameter <varname>log_statement</> has been
2470 changed so it can selectively log just database modification or
2471 data definition statements. Server configuration parameter
2472 <varname>log_duration</> now prints only when <varname>log_statement</>
2479 Server configuration parameter <varname>max_expr_depth</> parameter has
2480 been replaced with <varname>max_stack_depth</> which measures the
2481 physical stack size rather than the expression nesting depth. This
2482 helps prevent session termination due to stack overflow caused by
2483 recursive functions.
2489 The <function>length()</> function no longer counts trailing spaces in
2490 <type>CHAR(n)</> values.
2496 Casting an integer to <type>BIT(N)</> selects the rightmost N bits of the
2497 integer, not the leftmost N bits as before.
2503 Updating an element or slice of a NULL array value now produces
2504 a non-NULL array result, namely an array containing
2505 just the assigned-to positions.
2511 Syntax checking of array input values has been tightened up
2512 considerably. Junk that was previously allowed in odd places with
2513 odd results now causes an error. Empty-string element values
2514 must now be written as <literal>""</>, rather than writing nothing.
2515 Also changed behavior with respect to whitespace surrounding
2516 array elements: trailing whitespace is now ignored, for symmetry
2517 with leading whitespace (which has always been ignored).
2523 Overflow in integer arithmetic operations is now detected and
2524 reported as an error.
2530 The arithmetic operators associated with the single-byte
2531 <type>"char"</> data type have been removed.
2537 The <function>extract()</> function (also called
2538 <function>date_part</>) now returns the proper year for BC dates.
2539 It previously returned one less than the correct year. The
2540 function now also returns the proper values for millennium and
2547 <type>CIDR</> values now must have their non-masked bits be zero.
2548 For example, we no longer allow
2549 <literal>204.248.199.1/31</literal> as a <type>CIDR</> value. Such
2550 values should never have been accepted by
2551 <productname>PostgreSQL</productname> and will now be rejected.
2557 <command>EXECUTE</command> now returns a completion tag that
2558 matches the executed statement.
2564 <application>psql</>'s <command>\copy</> command now reads or
2565 writes to the query's <literal>stdin/stdout</>, rather than
2566 <application>psql</>'s <literal>stdin/stdout</>. The previous
2567 behavior can be accessed via new
2568 <option>pstdin</>/<option>pstdout</> parameters.
2574 The JDBC client interface has been removed from the core
2575 distribution, and is now hosted at <ulink url=
2576 "http://jdbc.postgresql.org"></ulink>.
2582 The Tcl client interface has also been removed. There are several
2583 Tcl interfaces now hosted at <ulink url=
2584 "http://gborg.postgresql.org"></ulink>.
2590 The server now uses its own time zone database, rather than the
2591 one supplied by the operating system. This will provide consistent
2592 behavior across all platforms. In most cases, there should be
2593 little noticeable difference in time zone behavior, except that
2594 the time zone names used by <command>SET</>/<command>SHOW</>
2595 <varname>TimeZone</> may
2596 be different from what your platform provides.
2602 <application>Configure</>'s threading option no longer requires
2603 users to run tests or edit configuration files; threading options
2604 are now detected automatically.
2610 Now that tablespaces have been implemented,
2611 <application>initlocation</> has been removed.
2617 The API for user-defined GiST indexes has been changed. The
2618 Union and PickSplit methods are now passed a pointer to a
2619 special <structname>GistEntryVector</structname> structure,
2620 rather than a <type>bytea</type>.
2628 <title>Deprecated Features</title>
2631 Some aspects of <productname>PostgreSQL</productname>'s behavior
2632 have been determined to be suboptimal. For the sake of backward
2633 compatibility these have not been removed in 8.0, but they are
2634 considered deprecated and will be removed in the next major
2641 The 8.1 release will remove the <function>to_char()</> function
2648 The server now warns of empty strings passed to
2649 <type>oid</type>/<type>float4</type>/<type>float8</type> data
2650 types, but continues to interpret them as zeroes as before.
2651 In the next major release, empty strings will be considered
2652 invalid input for these data types.
2658 By default, tables in <productname>PostgreSQL</productname> 8.0
2659 and earlier are created with <type>OID</>s. In the next release,
2660 this will <emphasis>not</emphasis> be the case: to create a table
2661 that contains <type>OID</>s, the <option>WITH OIDS</> clause must
2662 be specified or the <varname>default_with_oids</varname>
2663 configuration parameter must be set. Users are encouraged to
2664 explicitly specify <option>WITH OIDS</> if their tables
2665 require OIDs for compatibility with future releases of
2666 <productname>PostgreSQL</productname>.
2674 <title>Changes</title>
2677 Below you will find a detailed account of the changes between
2678 release 8.0 and the previous major release.
2682 <title>Performance Improvements</title>
2687 Support cross-data-type index usage (Tom)
2690 Before this change, many queries would not use an index if the data
2691 types did not match exactly. This improvement makes index usage more
2692 intuitive and consistent.
2698 New buffer replacement strategy that improves caching (Jan)
2701 Prior releases used a least-recently-used (LRU) cache to keep
2702 recently referenced pages in memory. The LRU algorithm
2703 did not consider the number of times a specific cache entry was
2704 accessed, so large table scans could force out useful cache pages.
2705 The new cache algorithm uses four separate lists to track most
2706 recently used and most frequently used cache pages and dynamically
2707 optimize their replacement based on the work load. This should
2708 lead to much more efficient use of the shared buffer cache.
2709 Administrators who have tested shared buffer sizes in the past
2710 should retest with this new cache replacement policy.
2716 Add subprocess to write dirty buffers periodically to reduce
2717 checkpoint writes (Jan)
2720 In previous releases, the checkpoint process, which runs every few
2721 minutes, would write all dirty buffers to the operating system's
2722 buffer cache then flush all dirty operating system buffers to
2723 disk. This resulted in a periodic spike in disk usage that often
2724 hurt performance. The new code uses a background writer to trickle
2725 disk writes at a steady pace so checkpoints have far fewer dirty
2726 pages to write to disk. Also, the new code does not issue a global
2727 <function>sync()</> call, but instead <function>fsync()</>s just
2728 the files written since the last checkpoint. This should improve
2729 performance and minimize degradation during checkpoints.
2735 Add ability to prolong vacuum to reduce performance impact (Jan)
2738 On busy systems, <command>VACUUM</command> performs many I/O
2739 requests which can hurt performance for other users. This
2740 release allows you to slow down <command>VACUUM</command> to
2741 reduce its impact on other users, though this increases the
2742 total duration of <command>VACUUM</command>.
2748 Improve B-tree index performance for duplicate keys (Dmitry Tkach, Tom)
2751 This improves the way indexes are scanned when many duplicate
2752 values exist in the index.
2758 Use dynamically-generated table size estimates while planning (Tom)
2761 Formerly the planner estimated table sizes using the values seen
2762 by the last <command>VACUUM</command> or <command>ANALYZE</command>,
2763 both as to physical table size (number of pages) and number of rows.
2764 Now, the current physical table size is obtained from the kernel,
2765 and the number of rows is estimated by multiplying the table size
2766 by the row density (rows per page) seen by the last
2767 <command>VACUUM</command> or <command>ANALYZE</command>. This should
2768 produce more reliable estimates in cases where the table size has
2769 changed significantly since the last housekeeping command.
2775 Improved index usage with <literal>OR</> clauses (Tom)
2778 This allows the optimizer to use indexes in statements with many OR
2779 clauses that would not have been indexed in the past. It can also use
2780 multi-column indexes where the first column is specified and the second
2781 column is part of an <literal>OR</> clause.
2787 Improve matching of partial index clauses (Tom)
2790 The server is now smarter about using partial indexes in queries
2791 involving complex <option>WHERE</> clauses.
2797 Improve performance of the GEQO optimizer (Tom)
2800 The GEQO optimizer is used to plan queries involving many tables (by
2801 default, twelve or more). This release speeds up the way queries are
2802 analyzed to decrease time spent in optimization.
2808 Miscellaneous optimizer improvements
2811 There is not room here to list all the minor improvements made, but
2812 numerous special cases work better than in prior releases.
2818 Improve lookup speed for C functions (Tom)
2821 This release uses a hash table to lookup information for dynamically
2822 loaded C functions. This improves their speed so they perform nearly as
2823 quickly as functions that are built into the server executable.
2829 Add type-specific <command>ANALYZE</command> statistics
2830 capability (Mark Cave-Ayland)
2833 This feature allows more flexibility in generating statistics
2834 for non-standard data types.
2840 <command>ANALYZE</command> now collects statistics for
2841 expression indexes (Tom)
2844 Expression indexes (also called functional indexes) allow users to
2845 index not just columns but the results of expressions and function
2846 calls. With this release, the optimizer can gather and use statistics
2847 about the contents of expression indexes. This will greatly improve
2848 the quality of planning for queries in which an expression index is
2855 New two-stage sampling method for <command>ANALYZE</command>
2859 This gives better statistics when the density of valid rows is very
2860 different in different regions of a table.
2866 Speed up <command>TRUNCATE</command> (Tom)
2869 This buys back some of the performance loss observed in 7.4, while still
2870 keeping <command>TRUNCATE</command> transaction-safe.
2879 <title>Server Changes</title>
2884 Add WAL file archiving and point-in-time recovery (Simon Riggs)
2890 Add tablespaces so admins can control disk layout (Gavin)
2896 Add a built-in log rotation program (Andreas Pflug)
2899 It is now possible to log server messages conveniently without
2900 relying on either <application>syslog</> or an external log
2907 Add new read-only server configuration parameters to show server
2908 compile-time settings: <varname>block_size</>,
2909 <varname>integer_datetimes</>, <varname>max_function_args</>,
2910 <varname>max_identifier_length</>, <varname>max_index_keys</> (Joe)
2916 Make quoting of <literal>sameuser</>, <literal>samegroup</>, and
2917 <literal>all</> remove special meaning of these terms in
2918 <filename>pg_hba.conf</> (Andrew)
2924 Use clearer IPv6 name <literal>::1/128</> for
2925 <literal>localhost</> in default <filename>pg_hba.conf</> (Andrew)
2931 Use CIDR format in <filename>pg_hba.conf</> examples (Andrew)
2937 Rename server configuration parameters <varname>SortMem</> and
2938 <varname>VacuumMem</> to <varname>work_mem</> and
2939 <varname>maintenance_work_mem</> (Old names still supported) (Tom)
2942 This change was made to clarify that bulk operations such as index and
2943 foreign key creation use <varname>maintenance_work_mem</>, while
2944 <varname>work_mem</> is for workspaces used during query execution.
2950 Allow logging of session disconnections using server configuration
2951 <varname>log_disconnections</> (Andrew)
2957 Add new server configuration parameter <varname>log_line_prefix</> to
2958 allow control of information emitted in each log line (Andrew)
2961 Available information includes user name, database name, remote IP
2962 address, and session start time.
2968 Remove server configuration parameters <varname>log_pid</>,
2969 <varname>log_timestamp</>, <varname>log_source_port</>; functionality
2970 superseded by <varname>log_line_prefix</> (Andrew)
2976 Replace the <varname>virtual_host</> and <varname>tcpip_socket</>
2977 parameters with a unified <varname>listen_addresses</> parameter
2981 <varname>virtual_host</> could only specify a single IP address to
2982 listen on. <varname>listen_addresses</> allows multiple addresses
2989 Listen on localhost by default, which eliminates the need for the
2990 <option>-i</> postmaster switch in many scenarios (Andrew)
2993 Listening on localhost (<literal>127.0.0.1</>) opens no new
2994 security holes but allows configurations like Windows and JDBC,
2995 which do not support local sockets, to work without special
3002 Remove <varname>syslog</> server configuration parameter, and add more
3003 logical <varname>log_destination</> variable to control log output
3010 Change server configuration parameter <varname>log_statement</> to take
3011 values <varname>all</>, <varname>mod</>, <varname>ddl</>, or
3012 <varname>none</> to select which queries are logged (Bruce)
3015 This allows administrators to log only data definition changes or
3016 only data modification statements.
3022 Some logging-related configuration parameters could formerly be adjusted
3023 by ordinary users, but only in the <quote>more verbose</> direction.
3024 They are now treated more strictly: only superusers can set them.
3025 However, a superuser may use <command>ALTER USER</> to provide per-user
3026 settings of these values for non-superusers. Also, it is now possible
3027 for superusers to set values of superuser-only configuration parameters
3028 via <literal>PGOPTIONS</>.
3034 Allow configuration files to be placed outside the data directory (mlw)
3037 By default, configuration files are kept in the cluster's top directory.
3038 With this addition, configuration files can be placed outside the
3039 data directory, easing administration.
3045 Plan prepared queries only when first executed so constants can be
3046 used for statistics (Oliver Jowett)
3049 Prepared statements plan queries once and execute them many
3050 times. While prepared queries avoid the overhead of re-planning
3051 on each use, the quality of the plan suffers from not knowing the exact
3052 parameters to be used in the query. In this release, planning of
3053 unnamed prepared statements is delayed until the first execution,
3054 and the actual parameter values of that execution are used as
3055 optimization hints. This allows use of out-of-line parameter passing
3056 without incurring a performance penalty.
3062 Allow <command>DECLARE CURSOR</command> to take parameters
3066 It is now useful to issue <command>DECLARE CURSOR</command> in a
3067 <function>Parse</> message with parameters. The parameter values
3068 sent at <function>Bind</> time will be substituted into the
3069 execution of the cursor's query.
3075 Fix hash joins and aggregates of <type>inet</type> and
3076 <type>cidr</type> data types (Tom)
3079 Release 7.4 handled hashing of mixed <type>inet</type> and
3080 <type>cidr</type> values incorrectly. (This bug did not exist
3081 in prior releases because they wouldn't try to hash either
3088 Make <varname>log_duration</> print only when <varname>log_statement</>
3089 prints the query (Ed L.)
3098 <title>Query Changes</title>
3103 Add savepoints (nested transactions) (Alvaro)
3109 Unsupported isolation levels are now accepted and promoted to the
3110 nearest supported level (Peter)
3113 The SQL specification states that if a database doesn't support a
3114 specific isolation level, it should use the next more restrictive level.
3115 This change complies with that recommendation.
3121 Allow <command>BEGIN WORK</command> to specify transaction
3122 isolation levels like <command>START TRANSACTION</command> does
3129 Fix table permission checking for cases in which rules generate
3130 a query type different from the originally submitted query (Tom)
3136 Implement dollar quoting to simplify single-quote usage (Andrew, Tom,
3140 In previous releases, because single quotes had to be used to
3141 quote a function's body, the use of single quotes inside the
3142 function text required use of two single quotes or other error-prone
3143 notations. With this release we add the ability to use "dollar
3144 quoting" to quote a block of text. The ability to use different
3145 quoting delimiters at different nesting levels greatly simplifies
3146 the task of quoting correctly, especially in complex functions.
3147 Dollar quoting can be used anywhere quoted text is needed.
3153 Make <literal>CASE val WHEN compval1 THEN ...</> evaluate <literal>val</> only once (Tom)
3156 <option>CASE</> no longer evaluates the tested expression multiple
3157 times. This has benefits when the expression is complex or is
3164 Test <option>HAVING</> before computing target list of an
3165 aggregate query (Tom)
3168 Fixes improper failure of cases such as <literal>SELECT SUM(win)/SUM(lose)
3169 ... GROUP BY ... HAVING SUM(lose) > 0</>. This should work but formerly
3170 could fail with divide-by-zero.
3176 Replace <varname>max_expr_depth</> parameter with
3177 <varname>max_stack_depth</> parameter, measured in kilobytes of stack
3181 This gives us a fairly bulletproof defense against crashing due to
3182 runaway recursive functions. Instead of measuring the depth of expression
3183 nesting, we now directly measure the size of the execution stack.
3189 Allow arbitrary row expressions (Tom)
3192 This release allows SQL expressions to contain arbitrary composite
3193 types, that is, row values. It also allows functions to more easily
3194 take rows as arguments and return row values.
3200 Allow <option>LIKE</>/<option>ILIKE</> to be used as the operator
3201 in row and subselect comparisons (Fabien Coelho)
3207 Avoid locale-specific case conversion of basic ASCII letters in
3208 identifiers and keywords (Tom)
3211 This solves the <quote>Turkish problem</> with mangling of words
3212 containing <literal>I</> and <literal>i</>. Folding of characters
3213 outside the 7-bit-ASCII set is still locale-aware.
3219 Improve syntax error reporting (Fabien, Tom)
3222 Syntax error reports are more useful than before.
3228 Change <command>EXECUTE</command> to return a completion tag
3229 matching the executed statement (Kris Jurka)
3232 Previous releases return an <command>EXECUTE</command> tag for
3233 any <command>EXECUTE</command> call. In this release, the tag
3234 returned will reflect the command executed.
3240 Avoid emitting <option>NATURAL CROSS JOIN</> in rule listings (Tom)
3243 Such a clause makes no logical sense, but in some cases the rule
3244 decompiler formerly produced this syntax.
3253 <title>Object Manipulation Changes</title>
3258 Add <command>COMMENT ON</> for casts, conversions, languages,
3259 operator classes, and large objects (Christopher)
3265 Add new server configuration parameter <varname>default_with_oids</> to
3266 control whether tables are created with <type>OID</>s by default (Neil)
3269 This allows administrators to control whether <command>CREATE
3270 TABLE</command> commands create tables with or without <type>OID</>
3271 columns by default. (Note: the current factory default setting for
3272 <varname>default_with_oids</> is <literal>TRUE</>, but the default
3273 will become <literal>FALSE</> in future releases.)
3279 Add <option>WITH</> / <option>WITHOUT OIDS</> clause to
3280 <command>CREATE TABLE AS</command> (Neil)
3286 Allow <command>ALTER TABLE DROP COLUMN</> to drop an <type>OID</>
3287 column (<command>ALTER TABLE SET WITHOUT OIDS</> still works)
3294 Allow composite types as table columns (Tom)
3300 Allow <command>ALTER ... ADD COLUMN</> with defaults and
3301 <option>NOT NULL</> constraints; works per SQL spec (Rod)
3304 It is now possible for <option>ADD COLUMN</> to create a column
3305 that is not initially filled with NULLs, but with a specified
3312 Add <command>ALTER COLUMN TYPE</> to change column's type (Rod)
3315 It is now possible to alter a column's data type without dropping
3316 and re-adding the column.
3322 Allow multiple <command>ALTER</> actions in a single <command>ALTER
3323 TABLE</command> command (Rod)
3326 This is particularly useful for <command>ALTER</> commands that
3327 rewrite the table (which include <option>ALTER COLUMN TYPE</> and
3328 <option>ADD COLUMN</> with a default). By grouping
3329 <command>ALTER</> commands together, the table need be rewritten
3336 Allow <command>ALTER TABLE</command> to add <type>SERIAL</type>
3340 This falls out from the new capability of specifying defaults for new
3347 Allow changing the owners of aggregates, conversions, databases,
3348 functions, operators, operator classes, schemas, types, and tablespaces
3349 (Christopher, Euler Taveira de Oliveira)
3352 Previously this required modifying the system tables directly.
3358 Allow temporary object creation to be limited to <option>SECURITY
3359 DEFINER</> functions (Sean Chittenden)
3365 Add <option>ALTER TABLE ... SET WITHOUT CLUSTER</> (Christopher)
3368 Prior to this release, there was no way to clear an auto-cluster
3369 specification except to modify the system tables.
3375 Constraint/Index/<type>SERIAL</> names are now
3376 <replaceable>table_column_type</>
3377 with numbers appended to guarantee uniqueness within the schema
3381 The SQL specification states that such names should be unique
3388 Add <function>pg_get_serial_sequence()</> to return a
3389 <type>SERIAL</> column's sequence name (Christopher)
3392 This allows automated scripts to reliably find the <type>SERIAL</>
3399 Warn when primary/foreign key data type mismatch requires costly lookup
3405 New <command>ALTER INDEX</> command to allow moving of indexes
3406 between tablespaces (Gavin)
3412 Make <command>ALTER TABLE OWNER</> change dependent sequence
3413 ownership too (Alvaro)
3423 <title>Utility Command Changes</title>
3428 Allow <command>CREATE SCHEMA</command> to create triggers,
3429 indexes, and sequences (Neil)
3435 Add <option>ALSO</> keyword to <command>CREATE RULE</command> (Fabien
3439 This allows <option>ALSO</> to be added to rule creation to contrast it with
3440 <option>INSTEAD</> rules.
3446 Add <option>NOWAIT</> option to <command>LOCK</command> (Tatsuo)
3449 This allows the <command>LOCK</command> command to fail if it
3450 would have to wait for the requested lock.
3456 Allow <command>COPY</command> to read and write
3457 comma-separated-value (CSV) files (Andrew, Bruce)
3463 Generate error if the <command>COPY</command> delimiter and NULL
3464 string conflict (Bruce)
3470 <command>GRANT</command>/<command>REVOKE</command> behavior
3471 follows the SQL spec more closely
3477 Avoid locking conflict between <command>CREATE INDEX</command>
3478 and <command>CHECKPOINT</command> (Tom)
3481 In 7.3 and 7.4, a long-running B-tree index build could block concurrent
3482 <command>CHECKPOINT</>s from completing, thereby causing WAL bloat because the
3483 WAL log could not be recycled.
3489 Database-wide <command>ANALYZE</command> does not hold locks
3493 This reduces the potential for deadlocks against other backends
3494 that want exclusive locks on tables. To get the benefit of this
3495 change, do not execute database-wide <command>ANALYZE</command>
3496 inside a transaction block (<command>BEGIN</command> block); it
3497 must be able to commit and start a new transaction for each
3504 <command>REINDEX</command> does not exclusively lock the index's
3505 parent table anymore
3508 The index itself is still exclusively locked, but readers of the
3509 table can continue if they are not using the particular index
3516 Erase MD5 user passwords when a user is renamed (Bruce)
3519 <productname>PostgreSQL</productname> uses the user name as salt
3520 when encrypting passwords via MD5. When a user's name is changed,
3521 the salt will no longer match the stored MD5 password, so the
3522 stored password becomes useless. In this release a notice is
3523 generated and the password is cleared. A new password must then
3524 be assigned if the user is to be able to log in with a password.
3530 New <application>pg_ctl</> <option>kill</> option for Windows (Andrew)
3533 Windows does not have a <literal>kill</> command to send signals to
3534 backends so this capability was added to <application>pg_ctl</>.
3540 Information schema improvements
3546 Add <option>--pwfile</> option to
3547 <application>initdb</application> so the initial password can be
3548 set by GUI tools (Magnus)
3554 Detect locale/encoding mismatch in
3555 <application>initdb</application> (Peter)
3561 Add <option>register</> command to <application>pg_ctl</> to
3562 register Windows operating system service (Dave Page)
3571 <title>Data Type and Function Changes</title>
3576 More complete support for composite types (row types) (Tom)
3579 Composite values can be used in many places where only scalar values
3586 Reject non-rectangular array values as erroneous (Joe)
3589 Formerly, <function>array_in</> would silently build a
3596 Overflow in integer arithmetic operations is now detected (Tom)
3602 The arithmetic operators associated with the single-byte
3603 <type>"char"</> data type have been removed.
3606 Formerly, the parser would select these operators in many situations
3607 where an <quote>unable to select an operator</> error would be more
3608 appropriate, such as <literal>null * null</>. If you actually want
3609 to do arithmetic on a <type>"char"</> column, you can cast it to
3616 Syntax checking of array input values considerably tightened up (Joe)
3619 Junk that was previously allowed in odd places with odd results
3620 now causes an <literal>ERROR</>, for example, non-whitespace
3621 after the closing right brace.
3627 Empty-string array element values must now be written as
3628 <literal>""</>, rather than writing nothing (Joe)
3631 Formerly, both ways of writing an empty-string element value were
3632 allowed, but now a quoted empty string is required. The case where
3633 nothing at all appears will probably be considered to be a NULL
3634 element value in some future release.
3640 Array element trailing whitespace is now ignored (Joe)
3643 Formerly leading whitespace was ignored, but trailing whitespace
3644 between an element value and the delimiter or right brace was
3645 significant. Now trailing whitespace is also ignored.
3651 Emit array values with explicit array bounds when lower bound is not one
3658 Accept <literal>YYYY-monthname-DD</> as a date string (Tom)
3664 Make <function>netmask</> and <function>hostmask</> functions
3665 return maximum-length mask length (Tom)
3671 Change factorial function to return <type>numeric</type> (Gavin)
3674 Returning <type>numeric</type> allows the factorial function to
3675 work for a wider range of input values.
3681 <function>to_char</>/<function>to_date()</> date conversion
3682 improvements (Kurt Roeckx, Fabien Coelho)
3688 Make <function>length()</> disregard trailing spaces in
3689 <type>CHAR(n)</> (Gavin)
3692 This change was made to improve consistency: trailing spaces are
3693 semantically insignificant in <type>CHAR(n)</> data, so they
3694 should not be counted by <function>length()</>.
3700 Warn about empty string being passed to
3701 <type>OID</>/<type>float4</>/<type>float8</> data types (Neil)
3704 8.1 will throw an error instead.
3710 Allow leading or trailing whitespace in
3711 <type>int2</>/<type>int4</>/<type>int8</>/<type>float4</>/<type>float8</>
3719 Better support for IEEE <literal>Infinity</> and <literal>NaN</>
3720 values in <type>float4</type>/<type>float8</type> (Neil)
3723 These should now work on all platforms that support IEEE-compliant
3724 floating point arithmetic.
3730 Add <option>week</> option to <function>date_trunc()</> (Robert Creager)
3736 Fix <function>to_char</function> for <literal>1 BC</>
3737 (previously it returned <literal>1 AD</>) (Bruce)
3743 Fix <function>date_part(year)</> for BC dates (previously it
3744 returned one less than the correct year) (Bruce)
3750 Fix <function>date_part()</> to return the proper millennium and
3751 century (Fabien Coelho)
3754 In previous versions, the century and millennium results had a wrong
3755 number and started in the wrong year, as compared to standard
3756 reckoning of such things.
3762 Add <function>ceiling()</> as an alias for <function>ceil()</>,
3763 and <function>power()</> as an alias for <function>pow()</> for
3764 standards compliance (Neil)
3770 Change <function>ln()</>, <function>log()</>,
3771 <function>power()</>, and <function>sqrt()</> to emit the correct
3772 <literal>SQLSTATE</> error codes for certain error conditions, as
3773 specified by SQL:2003 (Neil)
3779 Add <function>width_bucket()</> function as defined by SQL:2003 (Neil)
3785 Add <function>generate_series()</> functions to simplify working
3786 with numeric sets (Joe)
3792 Fix <function>upper/lower/initcap()</> functions to work with
3793 multibyte encodings (Tom)
3799 Add boolean and bitwise integer <option>AND</>/<option>OR</>
3800 aggregates (Fabien Coelho)
3806 New session information functions to return network addresses for client
3807 and server (Sean Chittenden)
3813 Add function to determine the area of a closed path (Sean Chittenden)
3819 Add function to send cancel request to other backends (Magnus)
3825 Add <type>interval</> plus <type>datetime</> operators (Tom)
3828 The reverse ordering, <type>datetime</> plus <type>interval</>,
3829 was already supported, but both are required by the SQL standard.
3835 Casting an integer to <type>BIT(N)</> selects the rightmost N bits
3840 In prior releases, the leftmost N bits were selected, but this was
3841 deemed unhelpful, not to mention inconsistent with casting from bit
3848 Require <type>CIDR</> values to have all non-masked bits be zero
3858 <title>Server-Side Language Changes</title>
3863 In <literal>READ COMMITTED</> serialization mode, volatile functions
3864 now see the results of concurrent transactions committed up to the
3865 beginning of each statement within the function, rather than up to the
3866 beginning of the interactive command that called the function.
3872 Functions declared <literal>STABLE</> or <literal>IMMUTABLE</> always
3873 use the snapshot of the calling query, and therefore do not see the
3874 effects of actions taken after the calling query starts, whether in
3875 their own transaction or other transactions. Such a function must be
3876 read-only, too, meaning that it cannot use any SQL commands other than
3877 <command>SELECT</>. There is a considerable performance gain from
3878 declaring a function <literal>STABLE</> or <literal>IMMUTABLE</>
3879 rather than <literal>VOLATILE</>.
3885 Non-deferred <option>AFTER</> triggers are now fired immediately
3886 after completion of the triggering query, rather than upon
3887 finishing the current interactive command. This makes a difference
3888 when the triggering query occurred within a function: the trigger
3889 is invoked before the function proceeds to its next operation. For
3890 example, if a function inserts a new row into a table, any
3891 non-deferred foreign key checks occur before proceeding with the
3898 Allow function parameters to be declared with names (Dennis Bjorklund)
3901 This allows better documentation of functions. Whether the names
3902 actually do anything depends on the specific function language
3909 Allow PL/pgSQL parameter names to be referenced in the function (Dennis Bjorklund)
3912 This basically creates an automatic alias for each named parameter.
3918 Do minimal syntax checking of PL/pgSQL functions at creation time (Tom)
3921 This allows us to catch simple syntax errors sooner.
3927 More support for composite types (row and record variables) in PL/pgSQL
3930 For example, it now works to pass a rowtype variable to another function
3931 as a single variable.
3937 Default values for PL/pgSQL variables can now reference previously
3944 Improve parsing of PL/pgSQL FOR loops (Tom)
3947 Parsing is now driven by presence of <literal>".."</> rather than
3948 data type of <option>FOR</> variable. This makes no difference for
3949 correct functions, but should result in more understandable error
3950 messages when a mistake is made.
3956 Major overhaul of PL/Perl server-side language (Command Prompt, Andrew Dunstan)
3962 In PL/Tcl, SPI commands are now run in subtransactions. If an error
3963 occurs, the subtransaction is cleaned up and the error is reported
3964 as an ordinary Tcl error, which can be trapped with <literal>catch</>.
3965 Formerly, it was not possible to catch such errors.
3971 Accept <command>ELSEIF</> in PL/pgSQL (Neil)
3974 Previously PL/pgSQL only allowed <command>ELSIF</>, but many people
3975 are accustomed to spelling this keyword <command>ELSEIF</>.
3984 <title><application>psql</> Changes</title>
3989 Improve <application>psql</> information display about database
3990 objects (Christopher)
3996 Allow <application>psql</> to display group membership in
3997 <command>\du</> and <command>\dg</> (Markus Bertheau)
4003 Prevent <application>psql</> <command>\dn</command> from showing
4004 temporary schemas (Bruce)
4010 Allow <application>psql</> to handle tilde user expansion for file
4017 Allow <application>psql</> to display fancy prompts, including
4018 color, via <application>readline</> (Reece Hart, Chet Ramey)
4024 Make <application>psql</> <command>\copy</> match <command>COPY</command> command syntax
4031 Show the location of syntax errors (Fabien Coelho, Tom)
4037 Add <command>CLUSTER</command> information to <application>psql</>
4038 <command>\d</> display
4045 Change <application>psql</> <command>\copy stdin/stdout</> to read
4046 from command input/output (Bruce)
4052 Add <option>pstdin</>/<option>pstdout</> to read from
4053 <application>psql</>'s <literal>stdin</>/<literal>stdout</> (Mark
4060 Add global <application>psql</> configuration file, <filename>psqlrc.sample</filename>
4064 This allows a central file where global <application>psql</> startup commands can
4071 Have <application>psql</> <command>\d+</> indicate if the table
4072 has an <type>OID</> column (Neil)
4078 On Windows, use binary mode in <application>psql</> when reading files so control-Z
4079 is not seen as end-of-file
4085 Have <command>\dn+</> show permissions and description for schemas (Dennis
4092 Improve tab completion support (Stefan Kaltenbrunn, Greg Sabino Mullane)
4098 Allow boolean settings to be set using upper or lower case (Michael Paesold)
4107 <title><application>pg_dump</> Changes</title>
4112 Use dependency information to improve the reliability of
4113 <application>pg_dump</> (Tom)
4116 This should solve the longstanding problems with related objects
4117 sometimes being dumped in the wrong order.
4123 Have <application>pg_dump</> output objects in alphabetical order if possible (Tom)
4126 This should make it easier to identify changes between
4133 Allow <application>pg_restore</> to ignore some SQL errors (Fabien Coelho)
4136 This makes <application>pg_restore</>'s behavior similar to the
4137 results of feeding a <application>pg_dump</> output script to
4138 <application>psql</>. In most cases, ignoring errors and plowing
4139 ahead is the most useful thing to do. Also added was a pg_restore
4140 option to give the old behavior of exiting on an error.
4146 <application>pg_restore</> <option>-l</> display now includes
4147 objects' schema names
4153 New begin/end markers in <application>pg_dump</> text output (Bruce)
4159 Add start/stop times for
4160 <application>pg_dump</>/<application>pg_dumpall</> in verbose mode
4167 Allow most <application>pg_dump</> options in
4168 <application>pg_dumpall</> (Christopher)
4174 Have <application>pg_dump</> use <command>ALTER OWNER</> rather
4175 than <command>SET SESSION AUTHORIZATION</> by default
4185 <title>libpq Changes</title>
4190 Make libpq's <option>SIGPIPE</> handling thread-safe (Bruce)
4196 Add <function>PQmbdsplen()</> which returns the display length
4197 of a character (Tatsuo)
4203 Add thread locking to <application>SSL</> and
4204 <application>Kerberos</> connections (Manfred Spraul)
4210 Allow <function>PQoidValue()</>, <function>PQcmdTuples()</>, and
4211 <function>PQoidStatus()</> to work on <command>EXECUTE</command>
4218 Add <function>PQserverVersion()</> to provide more convenient
4219 access to the server version number (Greg Sabino Mullane)
4225 Add <function>PQprepare/PQsendPrepared()</> functions to support
4226 preparing statements without necessarily specifying the data types
4227 of their parameters (Abhijit Menon-Sen)
4233 Many ECPG improvements, including <command>SET DESCRIPTOR</> (Michael)
4242 <title>Source Code Changes</title>
4247 Allow the database server to run natively on Windows (Claudio, Magnus, Andrew)
4253 Shell script commands converted to C versions for Windows support (Andrew)
4259 Create an extension makefile framework (Fabien Coelho, Peter)
4262 This simplifies the task of building extensions outside the original
4269 Support relocatable installations (Bruce)
4272 Directory paths for installed files (such as the
4273 <filename>/share</> directory) are now computed relative to the
4274 actual location of the executables, so that an installation tree
4275 can be moved to another place without reconfiguring and
4282 Use <option>--with-docdir</> to choose installation location of documentation; also
4283 allow <option>--infodir</> (Peter)
4289 Add <option>--without-docdir</> to prevent installation of documentation (Peter)
4295 Upgrade to <application>DocBook</> V4.2 SGML (Peter)
4301 New <literal>PostgreSQL</> <application>CVS</> tag (Marc)
4304 This was done to make it easier for organizations to manage their
4305 own copies of the <productname>PostgreSQL</productname>
4306 <application>CVS</> repository. File version stamps from the master
4307 repository will not get munged by checking into or out of a copied
4314 Clarify locking code (Manfred Koizar)
4320 Buffer manager cleanup (Neil)
4326 Decouple platform tests from CPU spinlock code (Bruce, Tom)
4332 Add inlined test-and-set code on PA-RISC for <application>gcc</>
4339 Improve i386 spinlock code (Manfred Spraul)
4345 Clean up spinlock assembly code to avoid warnings from newer
4346 <application>gcc</> releases (Tom)
4352 Remove JDBC from source tree; now a separate project
4358 Remove the libpgtcl client interface; now a separate project
4364 More accurately estimate memory and file descriptor usage (Tom)
4370 Improvements to the Mac OS X startup scripts (Ray A.)
4376 New <function>fsync()</> test program (Bruce)
4382 Major documentation improvements (Neil, Peter)
4388 Remove <application>pg_encoding</application>; not needed
4395 Remove <application>pg_id</application>; not needed anymore
4401 Remove <application>initlocation</application>; not needed
4408 Auto-detect thread flags (no more manual testing) (Bruce)
4414 Use Olson's public domain <application>timezone</> library (Magnus)
4420 With threading enabled, use thread flags on Unixware for
4421 backend executables too (Bruce)
4424 Unixware cannot mix threaded and non-threaded object files in the
4425 same executable, so everything must be compiled as threaded.
4431 <application>psql</> now uses a <application>flex</>-generated
4432 lexical analyzer to process command strings
4438 Reimplement the linked list data structure used throughout the
4442 This improves performance by allowing list append and length
4443 operations to be more efficient.
4449 Allow dynamically loaded modules to create their own server configuration
4450 parameters (Thomas Hallgren)
4456 New Brazilian version of FAQ (Euler Taveira de Oliveira)
4462 Add French FAQ (Guillaume Lelarge)
4468 New <application>pgevent</> for Windows logging
4474 Make libpq and ECPG build as proper shared libraries on OS X (Tom)
4483 <title>Contrib Changes</title>
4488 Overhaul of <filename>contrib/dblink</> (Joe)
4494 <filename>contrib/dbmirror</> improvements (Steven Singer)
4500 New <filename>contrib/xml2</> (John Gray, Torchbox)
4506 Updated <filename>contrib/mysql</filename>
4512 New version of <filename>contrib/btree_gist</> (Teodor)
4518 New <filename>contrib/trgm</>, trigram matching for
4519 <productname>PostgreSQL</productname> (Teodor)
4525 Many <filename>contrib/tsearch2</> improvements (Teodor)
4531 Add double metaphone to <filename>contrib/fuzzystrmatch</> (Andrew)
4537 Allow <filename>contrib/pg_autovacuum</> to run as a Windows service (Dave Page)
4543 Add functions to <filename>contrib/dbsize</> (Andreas Pflug)
4549 Removed <filename>contrib/pg_logger</>: obsoleted by integrated logging
4556 Removed <filename>contrib/rserv</>: obsoleted by various separate projects
4566 <sect1 id="release-7-4-8">
4567 <title>Release 7.4.8</title>
4570 <title>Release date</title>
4571 <simpara>2005-05-09</simpara>
4575 This release contains a variety of fixes from 7.4.7, including several
4576 security-related issues.
4580 <title>Migration to version 7.4.8</title>
4583 A dump/restore is not required for those running 7.4.X. However,
4584 it is one possible way of handling two significant security problems
4585 that have been found in the initial contents of 7.4.X system
4586 catalogs. A dump/initdb/reload sequence using 7.4.8's initdb will
4587 automatically correct these problems.
4591 The larger security problem is that the built-in character set encoding
4592 conversion functions can be invoked from SQL commands by unprivileged
4593 users, but the functions were not designed for such use and are not
4594 secure against malicious choices of arguments. The fix involves changing
4595 the declared parameter list of these functions so that they can no longer
4596 be invoked from SQL commands. (This does not affect their normal use
4597 by the encoding conversion machinery.)
4601 The lesser problem is that the <filename>contrib/tsearch2</> module
4602 creates several functions that are misdeclared to return
4603 <type>internal</> when they do not accept <type>internal</> arguments.
4604 This breaks type safety for all functions using <type>internal</>
4609 It is strongly recommended that all installations repair these errors,
4610 either by initdb or by following the manual repair procedures given
4611 below. The errors at least allow unprivileged database users to crash
4612 their server process, and may allow unprivileged users to gain the
4613 privileges of a database superuser.
4617 If you wish not to do an initdb, perform the following procedures instead.
4618 As the database superuser, do:
4622 UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
4623 WHERE pronamespace = 11 AND pronargs = 5
4624 AND proargtypes[2] = 'cstring'::regtype;
4625 -- The command should report having updated 90 rows;
4626 -- if not, rollback and investigate instead of committing!
4630 Next, if you have installed <filename>contrib/tsearch2</>, do
4634 UPDATE pg_proc SET proargtypes[0] = 'internal'::regtype
4636 'dex_init(text)'::regprocedure,
4637 'snb_en_init(text)'::regprocedure,
4638 'snb_ru_init(text)'::regprocedure,
4639 'spell_init(text)'::regprocedure,
4640 'syn_init(text)'::regprocedure
4642 -- The command should report having updated 5 rows;
4643 -- if not, rollback and investigate instead of committing!
4647 If this command fails with a message like <quote>function
4648 "dex_init(text)" does not exist</>, then either <filename>tsearch2</>
4649 is not installed in this database, or you already did the update.
4653 The above procedures must be carried out in <emphasis>each</> database
4654 of an installation, including <literal>template1</>, and ideally
4655 including <literal>template0</> as well. If you do not fix the
4656 template databases then any subsequently created databases will contain
4657 the same errors. <literal>template1</> can be fixed in the same way
4658 as any other database, but fixing <literal>template0</> requires
4659 additional steps. First, from any database issue
4661 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
4663 Next connect to <literal>template0</> and perform the above repair
4664 procedures. Finally, do
4666 -- re-freeze template0:
4668 -- and protect it against future alterations:
4669 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
4675 <title>Changes</title>
4678 <listitem><para>Change encoding function signature to prevent
4679 misuse</para></listitem>
4680 <listitem><para>Change <filename>contrib/tsearch2</> to avoid unsafe use of
4681 <type>INTERNAL</> function results</para></listitem>
4682 <listitem><para>Repair ancient race condition that allowed a transaction to be
4683 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
4684 than for other purposes</para>
4685 <para>This is an extremely serious bug since it could lead to apparent
4686 data inconsistencies being briefly visible to applications.</para></listitem>
4687 <listitem><para>Repair race condition between relation extension and
4689 <para>This could theoretically have caused loss of a page's worth of
4690 freshly-inserted data, although the scenario seems of very low probability.
4691 There are no known cases of it having caused more than an Assert failure.
4693 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
4695 The comparison code was wrong in the case where the
4696 <literal>--enable-integer-datetimes</> configuration switch had been used.
4697 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
4698 it will need to be <command>REINDEX</>ed after installing this update, because
4699 the fix corrects the sort order of column values.
4701 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
4702 <type>TIME WITH TIME ZONE</> values</para></listitem>
4703 <listitem><para>Fix mis-display of negative fractional seconds in
4704 <type>INTERVAL</> values</para>
4706 This error only occurred when the
4707 <literal>--enable-integer-datetimes</> configuration switch had been used.
4709 <listitem><para>Ensure operations done during backend shutdown are counted by
4710 statistics collector</para>
4712 This is expected to resolve reports of <application>pg_autovacuum</>
4713 not vacuuming the system catalogs often enough — it was not being
4714 told about catalog deletions caused by temporary table removal during
4717 <listitem><para>Additional buffer overrun checks in plpgsql
4718 (Neil)</para></listitem>
4719 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
4720 correctly (Neil)</para></listitem>
4721 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
4722 (Marko Kreen)</para></listitem>
4723 <listitem><para>Still more 64-bit fixes for
4724 <filename>contrib/intagg</></para></listitem>
4725 <listitem><para>Prevent incorrect optimization of functions returning
4726 <type>RECORD</></para></listitem>
4727 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
4728 month-related formats</para></listitem>
4729 <listitem><para>Prevent crash on <literal>COALESCE(NULL,NULL)</></para></listitem>
4730 <listitem><para>Fix <function>array_map</> to call PL functions correctly</para></listitem>
4731 <listitem><para>Fix permission checking in <command>ALTER DATABASE RENAME</></para></listitem>
4732 <listitem><para>Fix <command>ALTER LANGUAGE RENAME</></para></listitem>
4733 <listitem><para>Make <function>RemoveFromWaitQueue</> clean up after itself</para>
4735 This fixes a lock management error that would only be visible if a transaction
4736 was kicked out of a wait for a lock (typically by query cancel) and then the
4737 holder of the lock released it within a very narrow window.
4739 <listitem><para>Fix problem with untyped parameter appearing in
4740 <command>INSERT ... SELECT</></para></listitem>
4741 <listitem><para>Fix <command>CLUSTER</> failure after
4742 <command>ALTER TABLE SET WITHOUT OIDS</></para></listitem>
4748 <sect1 id="release-7-4-7">
4749 <title>Release 7.4.7</title>
4752 <title>Release date</title>
4753 <simpara>2005-01-31</simpara>
4757 This release contains a variety of fixes from 7.4.6, including several
4758 security-related issues.
4762 <title>Migration to version 7.4.7</title>
4765 A dump/restore is not required for those running 7.4.X.
4770 <title>Changes</title>
4773 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
4775 On platforms that will automatically execute initialization functions of a
4776 shared library (this includes at least Windows and ELF-based Unixen),
4777 <command>LOAD</> can be used to make the server execute arbitrary code.
4778 Thanks to NGS Software for reporting this.</para></listitem>
4779 <listitem><para>Check that creator of an aggregate function has the right to
4780 execute the specified transition functions</para>
4782 This oversight made it possible to bypass denial of EXECUTE
4783 permission on a function.</para></listitem>
4784 <listitem><para>Fix security and 64-bit issues in
4785 contrib/intagg</para></listitem>
4786 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
4787 Jurka)</para></listitem>
4788 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
4789 many parameters (Neil)</para></listitem>
4790 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
4792 The result of the join was mistakenly supposed to be sorted the same as the
4793 left input. This could not only deliver mis-sorted output to the user, but
4794 in case of nested merge joins could give outright wrong answers.
4796 <listitem><para>Fix plperl for quote marks in tuple fields</para></listitem>
4797 <listitem><para>Fix display of negative intervals in SQL and GERMAN
4798 datestyles</para></listitem>
4799 <listitem><para>Make age(timestamptz) do calculation in local timezone not
4800 GMT</para></listitem>
4806 <sect1 id="release-7-4-6">
4807 <title>Release 7.4.6</title>
4810 <title>Release date</title>
4811 <simpara>2004-10-22</simpara>
4815 This release contains a variety of fixes from 7.4.5.
4820 <title>Migration to version 7.4.6</title>
4823 A dump/restore is not required for those running 7.4.X.
4828 <title>Changes</title>
4831 <listitem><para>Repair possible failure to update hint bits on disk</para>
4833 Under rare circumstances this oversight could lead to
4834 <quote>could not access transaction status</> failures, which qualifies
4835 it as a potential-data-loss bug.
4837 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
4839 Very large left joins using a hash join plan could fail to output unmatched
4840 left-side rows given just the right data distribution.
4842 <listitem><para>Disallow running <application>pg_ctl</> as root</para>
4844 This is to guard against any possible security issues.
4846 <listitem><para>Avoid using temp files in <filename>/tmp</> in <command>make_oidjoins_check</command></para>
4848 This has been reported as a security issue, though it's hardly worthy of
4849 concern since there is no reason for non-developers to use this script anyway.
4851 <listitem><para>Prevent forced backend shutdown from re-emitting prior command
4854 In rare cases, a client might think that its last command had succeeded when
4855 it really had been aborted by forced database shutdown.
4857 <listitem><para>Repair bug in <function>pg_stat_get_backend_idset</function></para>
4859 This could lead to misbehavior in some of the system-statistics views.
4861 <listitem><para>Fix small memory leak in postmaster</para></listitem>
4862 <listitem><para>Fix <quote>expected both swapped tables to have TOAST
4863 tables</> bug</para>
4865 This could arise in cases such as CLUSTER after ALTER TABLE DROP COLUMN.
4867 <listitem><para>Prevent <literal>pg_ctl restart</> from adding <literal>-D</> multiple times</para></listitem>
4868 <listitem><para>Fix problem with NULL values in GiST indexes</para></listitem>
4869 <listitem><para><literal>::</> is no longer interpreted as a variable in an
4870 ECPG prepare statement</para></listitem>
4876 <sect1 id="release-7-4-5">
4877 <title>Release 7.4.5</title>
4880 <title>Release date</title>
4881 <simpara>2004-08-18</simpara>
4885 This release contains one serious bug fix over 7.4.4.
4890 <title>Migration to version 7.4.5</title>
4893 A dump/restore is not required for those running 7.4.X.
4898 <title>Changes</title>
4901 <listitem><para>Repair possible crash during concurrent B-tree index insertions</para>
4903 This patch fixes a rare case in which concurrent insertions into a B-tree index
4904 could result in a server panic. No permanent damage would result, but it's
4905 still worth a re-release. The bug does not exist in pre-7.4 releases.
4912 <sect1 id="release-7-4-4">
4913 <title>Release 7.4.4</title>
4916 <title>Release date</title>
4917 <simpara>2004-08-16</simpara>
4921 This release contains a variety of fixes from 7.4.3.
4926 <title>Migration to version 7.4.4</title>
4929 A dump/restore is not required for those running 7.4.X.
4934 <title>Changes</title>
4937 <listitem><para>Prevent possible loss of committed transactions during crash</para>
4939 Due to insufficient interlocking between transaction commit and checkpointing,
4940 it was possible for transactions committed just before the most recent
4941 checkpoint to be lost, in whole or in part, following a database crash and
4942 restart. This is a serious bug that has existed
4943 since <productname>PostgreSQL</productname> 7.1.
4945 <listitem><para>Check HAVING restriction before evaluating result list of an
4946 aggregate plan</para></listitem>
4947 <listitem><para>Avoid crash when session's current user ID is deleted</para></listitem>
4948 <listitem><para>Fix hashed crosstab for zero-rows case (Joe)</para></listitem>
4949 <listitem><para>Force cache update after renaming a column in a foreign key</para></listitem>
4950 <listitem><para>Pretty-print UNION queries correctly</para></listitem>
4951 <listitem><para>Make psql handle <literal>\r\n</> newlines properly in COPY IN</para></listitem>
4952 <listitem><para><application>pg_dump</> handled ACLs with grant options incorrectly</para></listitem>
4953 <listitem><para>Fix thread support for OS X and Solaris</para></listitem>
4954 <listitem><para>Updated JDBC driver (build 215) with various fixes</para></listitem>
4955 <listitem><para>ECPG fixes</para></listitem>
4956 <listitem><para>Translation updates (various contributors)</para></listitem>
4962 <sect1 id="release-7-4-3">
4963 <title>Release 7.4.3</title>
4966 <title>Release date</title>
4967 <simpara>2004-06-14</simpara>
4971 This release contains a variety of fixes from 7.4.2.
4976 <title>Migration to version 7.4.3</title>
4979 A dump/restore is not required for those running 7.4.X.
4984 <title>Changes</title>
4987 <listitem><para>Fix temporary memory leak when using non-hashed aggregates (Tom)</para></listitem>
4988 <listitem><para>ECPG fixes, including some for Informix compatibility (Michael)</para></listitem>
4989 <listitem><para>Fixes for compiling with thread-safety, particularly Solaris (Bruce)</para></listitem>
4990 <listitem><para>Fix error in COPY IN termination when using the old network protocol (ljb)</para></listitem>
4991 <listitem><para>Several important fixes in pg_autovacuum, including fixes for
4992 large tables, unsigned oids, stability, temp tables, and debug mode
4993 (Matthew T. O'Connor)</para></listitem>
4994 <listitem><para>Fix problem with reading tar-format dumps on NetBSD and BSD/OS (Bruce)</para></listitem>
4995 <listitem><para>Several JDBC fixes</para></listitem>
4996 <listitem><para>Fix ALTER SEQUENCE RESTART where last_value equals the restart value (Tom)</para></listitem>
4997 <listitem><para>Repair failure to recalculate nested sub-selects (Tom)</para></listitem>
4998 <listitem><para>Fix problems with non-constant expressions in LIMIT/OFFSET</para></listitem>
4999 <listitem><para>Support FULL JOIN with no join clause, such as X FULL JOIN Y ON TRUE (Tom)</para></listitem>
5000 <listitem><para>Fix another zero-column table bug (Tom)</para></listitem>
5001 <listitem><para>Improve handling of non-qualified identifiers in GROUP BY clauses in sub-selects (Tom)</para>
5003 Select-list aliases within the sub-select will now take precedence over
5004 names from outer query levels.
5006 <listitem><para>Do not generate <quote>NATURAL CROSS JOIN</> when decompiling rules (Tom)</para></listitem>
5007 <listitem><para>Add checks for invalid field length in binary COPY (Tom)</para>
5009 This fixes a difficult-to-exploit security hole.
5011 <listitem><para>Avoid locking conflict between <command>ANALYZE</command> and <command>LISTEN</command>/<command>NOTIFY</command></para></listitem>
5012 <listitem><para>Numerous translation updates (various contributors)</para></listitem>
5018 <sect1 id="release-7-4-2">
5019 <title>Release 7.4.2</title>
5022 <title>Release date</title>
5023 <simpara>2004-03-08</simpara>
5027 This release contains a variety of fixes from 7.4.1.
5032 <title>Migration to version 7.4.2</title>
5035 A dump/restore is not required for those running 7.4.X. However,
5036 it may be advisable as the easiest method of incorporating fixes for
5037 two errors that have been found in the initial contents of 7.4.X system
5038 catalogs. A dump/initdb/reload sequence using 7.4.2's initdb will
5039 automatically correct these problems.
5043 The more severe of the two errors is that data type <type>anyarray</>
5044 has the wrong alignment label; this is a problem because the
5045 <structname>pg_statistic</> system catalog uses <type>anyarray</>
5046 columns. The mislabeling can cause planner misestimations and even
5047 crashes when planning queries that involve <literal>WHERE</> clauses on
5048 double-aligned columns (such as <type>float8</> and <type>timestamp</>).
5049 It is strongly recommended that all installations repair this error,
5050 either by initdb or by following the manual repair procedure given
5055 The lesser error is that the system view <structname>pg_settings</>
5056 ought to be marked as having public update access, to allow
5057 <literal>UPDATE pg_settings</> to be used as a substitute for
5058 <command>SET</>. This can also be fixed either by initdb or manually,
5059 but it is not necessary to fix unless you want to use <literal>UPDATE
5064 If you wish not to do an initdb, the following procedure will work
5065 for fixing <structname>pg_statistic</>. As the database superuser,
5069 -- clear out old data in pg_statistic:
5070 DELETE FROM pg_statistic;
5071 VACUUM pg_statistic;
5072 -- this should update 1 row:
5073 UPDATE pg_type SET typalign = 'd' WHERE oid = 2277;
5074 -- this should update 6 rows:
5075 UPDATE pg_attribute SET attalign = 'd' WHERE atttypid = 2277;
5077 -- At this point you MUST start a fresh backend to avoid a crash!
5079 -- repopulate pg_statistic:
5083 This can be done in a live database, but beware that all backends
5084 running in the altered database must be restarted before it is safe to
5085 repopulate <structname>pg_statistic</>.
5089 To repair the <structname>pg_settings</> error, simply do:
5091 GRANT SELECT, UPDATE ON pg_settings TO PUBLIC;
5096 The above procedures must be carried out in <emphasis>each</> database
5097 of an installation, including <literal>template1</>, and ideally
5098 including <literal>template0</> as well. If you do not fix the
5099 template databases then any subsequently created databases will contain
5100 the same errors. <literal>template1</> can be fixed in the same way
5101 as any other database, but fixing <literal>template0</> requires
5102 additional steps. First, from any database issue
5104 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
5106 Next connect to <literal>template0</> and perform the above repair
5107 procedures. Finally, do
5109 -- re-freeze template0:
5111 -- and protect it against future alterations:
5112 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
5118 <title>Changes</title>
5121 Release 7.4.2 incorporates all the fixes included in release 7.3.6,
5122 plus the following fixes:
5126 <listitem><para>Fix <structname>pg_statistics</> alignment bug that could crash optimizer</para>
5127 <para>See above for details about this problem.</para></listitem>
5128 <listitem><para>Allow non-super users to update <structname>pg_settings</></para></listitem>
5129 <listitem><para>Fix several optimizer bugs, most of which led to
5130 <quote>variable not found in subplan target lists</> errors</para></listitem>
5131 <listitem><para>Avoid out-of-memory failure during startup of large multiple
5132 index scan</para></listitem>
5133 <listitem><para>Fix multibyte problem that could lead to <quote>out of
5134 memory</> error during <command>COPY IN</></para></listitem>
5135 <listitem><para>Fix problems with <command>SELECT INTO</> / <command>CREATE
5136 TABLE AS</> from tables without OIDs</para></listitem>
5137 <listitem><para>Fix problems with <filename>alter_table</> regression test
5138 during parallel testing</para></listitem>
5139 <listitem><para>Fix problems with hitting open file limit, especially on OS X (Tom)</para></listitem>
5140 <listitem><para>Partial fix for Turkish-locale issues</para>
5141 <para>initdb will succeed now in Turkish locale, but there are still some
5142 inconveniences associated with the <literal>i/I</> problem.</para></listitem>
5143 <listitem><para>Make pg_dump set client encoding on restore</para></listitem>
5144 <listitem><para>Other minor pg_dump fixes</para></listitem>
5145 <listitem><para>Allow ecpg to again use C keywords as column names (Michael)</para></listitem>
5146 <listitem><para>Added ecpg <literal>WHENEVER NOT_FOUND</> to
5147 <literal>SELECT/INSERT/UPDATE/DELETE</> (Michael)</para></listitem>
5148 <listitem><para>Fix ecpg crash for queries calling set-returning functions (Michael)</para></listitem>
5149 <listitem><para>Various other ecpg fixes (Michael)</para></listitem>
5150 <listitem><para>Fixes for Borland compiler</para></listitem>
5151 <listitem><para>Thread build improvements (Bruce)</para></listitem>
5152 <listitem><para>Various other build fixes</para></listitem>
5153 <listitem><para>Various JDBC fixes</para></listitem>
5159 <sect1 id="release-7-4-1">
5160 <title>Release 7.4.1</title>
5163 <title>Release date</title>
5164 <simpara>2003-12-22</simpara>
5168 This release contains a variety of fixes from 7.4.
5173 <title>Migration to version 7.4.1</title>
5176 A dump/restore is <emphasis>not</emphasis> required for those
5181 If you want to install the fixes in the information schema
5182 you need to reload it into the database.
5183 This is either accomplished by initializing a new cluster
5184 by running <command>initdb</command>, or by running the following
5185 sequence of SQL commands in each database (ideally including
5186 <literal>template1</literal>) as a superuser in
5187 <application>psql</application>, after installing the new release:
5189 DROP SCHEMA information_schema CASCADE;
5190 \i /usr/local/pgsql/share/information_schema.sql
5192 Substitute your installation path in the second command.
5198 <title>Changes</title>
5201 <listitem><para>Fixed bug in <command>CREATE SCHEMA</command> parsing in ECPG (Michael)</para></listitem>
5202 <listitem><para>Fix compile error when <option>--enable-thread-safety</option> and <option>--with-perl</option> are used together (Peter)</para></listitem>
5203 <listitem><para>Fix for subqueries that used hash joins (Tom)</para>
5205 Certain subqueries that used hash joins would crash because of
5206 improperly shared structures.
5208 <listitem><para>Fix free space map compaction bug (Tom)</para>
5210 This fixes a bug where compaction of the free space map could lead
5211 to a database server shutdown.
5214 <listitem><para>Fix for Borland compiler build of libpq (Bruce)</para></listitem>
5215 <listitem><para>Fix <function>netmask()</function> and <function>hostmask()</function> to return the maximum-length masklen (Tom)</para>
5217 Fix these functions to return values consistent with pre-7.4
5221 <listitem><para>Several <filename>contrib/pg_autovacuum</filename> fixes</para>
5223 Fixes include improper variable initialization, missing vacuum after
5224 <command>TRUNCATE</command>, and duration computation overflow for long vacuums.
5227 <listitem><para>Allow compile of <filename>contrib/cube</filename> under Cygwin (Jason Tishler)</para></listitem>
5228 <listitem><para>Fix Solaris use of password file when no passwords are defined (Tom)</para>
5230 Fix crash on Solaris caused by use of any type of password
5231 authentication when no passwords were defined.
5234 <listitem><para>JDBC fix for thread problems, other fixes</para></listitem>
5235 <listitem><para>Fix for <type>bytea</type> index lookups (Joe)</para></listitem>
5236 <listitem><para>Fix information schema for bit data types (Peter)</para></listitem>
5237 <listitem><para>Force zero_damaged_pages to be on during recovery from WAL</para></listitem>
5238 <listitem><para>Prevent some obscure cases of <quote>variable not in subplan target lists</quote></para></listitem>
5239 <listitem><para>Make <function>PQescapeBytea</function> and <function>byteaout</function> consistent with each other (Joe)</para></listitem>
5240 <listitem><para>Escape <type>bytea</type> output for bytes > 0x7e(Joe)</para>
5242 If different client encodings are used for <type>bytea</type> output and input, it
5243 is possible for <type>bytea</type> values to be corrupted by the differing
5244 encodings. This fix escapes all bytes that might be affected.
5247 <listitem><para>Added missing <function>SPI_finish()</function> calls to dblink's <function>get_tuple_of_interest()</function> (Joe)</para></listitem>
5248 <listitem><para>New Czech FAQ</para></listitem>
5249 <listitem><para>Fix information schema view <literal>constraint_column_usage</literal> for foreign keys (Peter)</para></listitem>
5250 <listitem><para>ECPG fixes (Michael)</para></listitem>
5251 <listitem><para>Fix bug with multiple <literal>IN</literal> subqueries and joins in the subqueries (Tom)</para></listitem>
5252 <listitem><para>Alllow <literal>COUNT('x')</literal> to work (Tom)</para></listitem>
5253 <listitem><para>Install ECPG include files for Informix compatibility into separate directory (Peter)</para>
5255 Some names of ECPG include files for Informix compatibility conflicted with operating system include files.
5256 By installing them in their own directory, name conflicts have been reduced.
5259 <listitem><para>Fix SSL memory leak (Neil)</para>
5261 This release fixes a bug in 7.4 where SSL didn't free all memory it allocated.
5264 <listitem><para>Prevent <filename>pg_service.conf</filename> from using service name as default dbname (Bruce)</para></listitem>
5265 <listitem><para>Fix local ident authentication on FreeBSD (Tom)</para></listitem>
5271 <sect1 id="release-7-4">
5272 <title>Release 7.4</title>
5275 <title>Release date</title>
5276 <simpara>2003-11-17</simpara>
5280 <title>Overview</title>
5283 Major changes in this release:
5289 <literal>IN</literal> / <literal>NOT IN</literal> subqueries are
5290 now much more efficient
5295 In previous releases, <literal>IN</literal>/<literal>NOT
5296 IN</literal> subqueries were joined to the upper query by
5297 sequentially scanning the subquery looking for a match. The
5298 7.4 code uses the same sophisticated techniques used by
5299 ordinary joins and so is much faster. An
5300 <literal>IN</literal> will now usually be as fast as or faster
5301 than an equivalent <literal>EXISTS</literal> subquery; this
5302 reverses the conventional wisdom that applied to previous
5310 Improved <literal>GROUP BY</literal> processing by using hash buckets
5315 In previous releases, rows to be grouped had to be sorted
5316 first. The 7.4 code can do <literal>GROUP BY</literal>
5317 without sorting, by accumulating results into a hash table
5318 with one entry per group. It will still use the sort
5319 technique, however, if the hash table is estimated to be too
5320 large to fit in <varname>sort_mem</>.
5327 New multikey hash join capability
5332 In previous releases, hash joins could only occur on single
5333 keys. This release allows multicolumn hash joins.
5340 Queries using the explicit <literal>JOIN</literal> syntax are
5341 now better optimized
5346 Prior releases evaluated queries using the explicit
5347 <literal>JOIN</literal> syntax only in the order implied by
5348 the syntax. 7.4 allows full optimization of these queries,
5349 meaning the optimizer considers all possible join orderings
5350 and chooses the most efficient. Outer joins, however, must
5351 still follow the declared ordering.
5358 Faster and more powerful regular expression code
5363 The entire regular expression module has been replaced with a
5364 new version by Henry Spencer, originally written for Tcl. The
5365 code greatly improves performance and supports several flavors
5366 of regular expressions.
5373 Function-inlining for simple SQL functions
5378 Simple SQL functions can now be inlined by including their SQL
5379 in the main query. This improves performance by eliminating
5380 per-call overhead. That means simple SQL functions now
5388 Full support for IPv6 connections and IPv6 address data types
5393 Previous releases allowed only IPv4 connections, and the IP
5394 data types only supported IPv4 addresses. This release adds
5395 full IPv6 support in both of these areas.
5402 Major improvements in SSL performance and reliability
5407 Several people very familiar with the SSL API have overhauled
5408 our SSL code to improve SSL key negotiation and error
5416 Make free space map efficiently reuse empty index pages,
5417 and other free space management improvements
5422 In previous releases, B-tree index pages that were left empty
5423 because of deleted rows could only be reused by rows with
5424 index values similar to the rows originally indexed on that
5425 page. In 7.4, <command>VACUUM</command> records empty index
5426 pages and allows them to be reused for any future index rows.
5433 SQL-standard information schema
5438 The information schema provides a standardized and stable way
5439 to access information about the schema objects defined in a
5447 Cursors conform more closely to the SQL standard
5452 The commands <command>FETCH</command> and
5453 <command>MOVE</command> have been overhauled to conform more
5454 closely to the SQL standard.
5461 Cursors can exist outside transactions
5466 These cursors are also called holdable cursors.
5473 New client-to-server protocol
5478 The new protocol adds error codes, more status information,
5479 faster startup, better support for binary data transmission,
5480 parameter values separated from SQL commands, prepared
5481 statements available at the protocol level, and cleaner
5482 recovery from <command>COPY</command> failures. The older
5483 protocol is still supported by both server and clients.
5490 <application>libpq</application> and
5491 <application>ECPG</application> applications are now fully
5497 While previous <application>libpq</application> releases
5498 already supported threads, this release improves thread safety
5499 by fixing some non-thread-safe code that was used during
5500 database connection startup. The <command>configure</command>
5501 option <option>--enable-thread-safety</option> must be used to
5502 enable this feature.
5509 New version of full-text indexing
5514 A new full-text indexing suite is available in
5515 <filename>contrib/tsearch2</filename>.
5527 The new autovacuum tool in
5528 <filename>contrib/autovacuum</filename> monitors the database
5529 statistics tables for
5530 <command>INSERT</command>/<command>UPDATE</command>/<command>DELETE</command>
5531 activity and automatically vacuums tables when needed.
5538 Array handling has been improved and moved into the server core
5543 Many array limitations have been removed, and arrays behave
5544 more like fully-supported data types.
5553 <title>Migration to version 7.4</title>
5556 A dump/restore using <application>pg_dump</application> is
5557 required for those wishing to migrate data from any previous
5562 Observe the following incompatibilities:
5568 The server-side autocommit setting was removed and
5569 reimplemented in client applications and languages.
5570 Server-side autocommit was causing too many problems with
5571 languages and applications that wanted to control their own
5572 autocommit behavior, so autocommit was removed from the server
5573 and added to individual client APIs as appropriate.
5579 Error message wording has changed substantially in this
5580 release. Significant effort was invested to make the messages
5581 more consistent and user-oriented. If your applications try to
5582 detect different error conditions by parsing the error message,
5583 you are strongly encouraged to use the new error code facility instead.
5589 Inner joins using the explicit <literal>JOIN</literal> syntax
5590 may behave differently because they are now better
5597 A number of server configuration parameters have been renamed
5598 for clarity, primarily those related to
5605 <literal>FETCH 0</literal> or <literal>MOVE 0</literal> now
5606 does nothing. In prior releases, <literal>FETCH 0</literal>
5607 would fetch all remaining rows, and <literal>MOVE 0</literal>
5608 would move to the end of the cursor.
5614 <command>FETCH</command> and <command>MOVE</command> now return
5615 the actual number of rows fetched/moved, or zero if at the
5616 beginning/end of the cursor. Prior releases would return the
5617 row count passed to the command, not the number of rows
5618 actually fetched or moved.
5624 <command>COPY</command> now can process files that use
5625 carriage-return or carriage-return/line-feed end-of-line
5626 sequences. Literal carriage-returns and line-feeds are no
5627 longer accepted in data values; use <literal>\r</literal> and
5628 <literal>\n</literal> instead.
5634 Trailing spaces are now trimmed when converting from type
5635 <type>char(<replaceable>n</>)</type> to
5636 <type>varchar(<replaceable>n</>)</type> or <type>text</type>.
5637 This is what most people always expected to happen anyway.
5643 The data type <type>float(<replaceable>p</>)</type> now
5644 measures <replaceable>p</> in binary digits, not decimal
5645 digits. The new behavior follows the SQL standard.
5651 Ambiguous date values now must match the ordering specified by
5652 the <varname>datestyle</varname> setting. In prior releases, a
5653 date specification of <literal>10/20/03</> was interpreted as a
5654 date in October even if <varname>datestyle</> specified that
5655 the day should be first. 7.4 will throw an error if a date
5656 specification is invalid for the current setting of
5657 <varname>datestyle</>.
5663 The functions <function>oidrand</function>,
5664 <function>oidsrand</function>, and
5665 <function>userfntest</function> have been removed. These
5666 functions were determined to be no longer useful.
5672 String literals specifying time-varying date/time values, such
5673 as <literal>'now'</literal> or <literal>'today'</literal> will
5674 no longer work as expected in column default expressions; they
5675 now cause the time of the table creation to be the default, not
5676 the time of the insertion. Functions such as
5677 <function>now()</>, <function>current_timestamp</>, or
5678 <function>current_date</function> should be used instead.
5682 In previous releases, there was special code so that strings
5683 such as <literal>'now'</literal> were interpreted at
5684 <command>INSERT</> time and not at table creation time, but
5685 this work around didn't cover all cases. Release 7.4 now
5686 requires that defaults be defined properly using functions such
5687 as <function>now()</> or <function>current_timestamp</>. These
5688 will work in all situations.
5694 The dollar sign (<literal>$</>) is no longer allowed in
5695 operator names. It can instead be a non-first character in
5696 identifiers. This was done to improve compatibility with other
5697 database systems, and to avoid syntax problems when parameter
5698 placeholders (<literal>$<replaceable>n</></>) are written
5699 adjacent to operators.
5707 <title>Changes</title>
5710 Below you will find a detailed account of the changes between
5711 release 7.4 and the previous major release.
5715 <title>Server Operation Changes</title>
5720 Allow IPv6 server connections (Nigel Kukard, Johan Jordaan,
5721 Bruce, Tom, Kurt Roeckx, Andrew Dunstan)
5727 Fix SSL to handle errors cleanly (Nathan Mueller)
5730 In prior releases, certain SSL API error reports were not
5731 handled correctly. This release fixes those problems.
5737 SSL protocol security and performance improvements (Sean Chittenden)
5740 SSL key renegotiation was happening too frequently, causing poor
5741 SSL performance. Also, initial key handling was improved.
5747 Print lock information when a deadlock is detected (Tom)
5750 This allows easier debugging of deadlock situations.
5756 Update <filename>/tmp</filename> socket modification times
5757 regularly to avoid their removal (Tom)
5760 This should help prevent <filename>/tmp</filename> directory
5761 cleaner administration scripts from removing server socket
5766 <listitem><para>Enable PAM for Mac OS X (Aaron Hillegass)</para></listitem>
5769 <para>Make B-tree indexes fully WAL-safe (Tom)</para>
5771 In prior releases, under certain rare cases, a server crash
5772 could cause B-tree indexes to become corrupt. This release
5773 removes those last few rare cases.
5777 <listitem><para>Allow B-tree index compaction and empty page reuse (Tom)</para></listitem>
5781 Fix inconsistent index lookups during split of first root page (Tom)
5784 In prior releases, when a single-page index split into two
5785 pages, there was a brief period when another database session
5786 could miss seeing an index entry. This release fixes that rare
5791 <listitem><para>Improve free space map allocation logic (Tom)</para></listitem>
5794 <para>Preserve free space information between server restarts (Tom)</para>
5796 In prior releases, the free space map was not saved when the
5797 postmaster was stopped, so newly started servers had no free
5798 space information. This release saves the free space map, and
5799 reloads it when the server is restarted.
5803 <listitem><para>Add start time to <literal>pg_stat_activity</literal> (Neil)</para></listitem>
5804 <listitem><para>New code to detect corrupt disk pages; erase with <varname>zero_damaged_pages</varname> (Tom)</para></listitem>
5805 <listitem><para>New client/server protocol: faster, no username length limit, allow clean exit from <command>COPY</command> (Tom)</para></listitem>
5806 <listitem><para>Add transaction status, table ID, column ID to client/server protocol (Tom)</para></listitem>
5807 <listitem><para>Add binary I/O to client/server protocol (Tom)</para></listitem>
5808 <listitem><para>Remove autocommit server setting; move to client applications (Tom)</para></listitem>
5809 <listitem><para>New error message wording, error codes, and three levels of error detail (Tom, Joe, Peter)</para></listitem>
5814 <title>Performance Improvements</title>
5817 <listitem><para>Add hashing for <literal>GROUP BY</literal> aggregates (Tom)</para></listitem>
5818 <listitem><para>Make nested-loop joins be smarter about multicolumn indexes (Tom)</para></listitem>
5819 <listitem><para>Allow multikey hash joins (Tom)</para></listitem>
5820 <listitem><para>Improve constant folding (Tom)</para></listitem>
5821 <listitem><para>Add ability to inline simple SQL functions (Tom)</para></listitem>
5824 <para>Reduce memory usage for queries using complex functions (Tom)</para>
5826 In prior releases, functions returning allocated memory would
5827 not free it until the query completed. This release allows the
5828 freeing of function-allocated memory when the function call
5829 completes, reducing the total memory used by functions.
5834 <para>Improve GEQO optimizer performance (Tom)</para>
5836 This release fixes several inefficiencies in the way the GEQO optimizer
5837 manages potential query paths.
5843 Allow <literal>IN</>/<literal>NOT IN</> to be handled via hash
5850 Improve <literal>NOT IN (<replaceable>subquery</>)</literal>
5857 Allow most <literal>IN</literal> subqueries to be processed as
5864 Pattern matching operations can use indexes regardless of
5868 There is no way for non-ASCII locales to use the standard
5869 indexes for <literal>LIKE</literal> comparisons. This release
5870 adds a way to create a special index for
5871 <literal>LIKE</literal>.
5876 <para>Allow the postmaster to preload libraries using <varname>preload_libraries</varname> (Joe)</para>
5878 For shared libraries that require a long time to load, this
5879 option is available so the library can be preloaded in the
5880 postmaster and inherited by all database sessions.
5886 Improve optimizer cost computations, particularly for subqueries (Tom)
5892 Avoid sort when subquery <literal>ORDER BY</literal> matches upper query (Tom)
5898 Deduce that <literal>WHERE a.x = b.y AND b.y = 42</literal> also
5899 means <literal>a.x = 42</literal> (Tom)
5905 Allow hash/merge joins on complex joins (Tom)
5911 Allow hash joins for more data types (Tom)
5917 Allow join optimization of explicit inner joins, disable with
5918 <varname>join_collapse_limit</varname> (Tom)
5924 Add parameter <varname>from_collapse_limit</varname> to control
5925 conversion of subqueries to joins (Tom)
5931 Use faster and more powerful regular expression code from Tcl
5932 (Henry Spencer, Tom)
5938 Use bit-mapped relation sets in the optimizer (Tom)
5943 <para>Improve connection startup time (Tom)</para>
5945 The new client/server protocol requires fewer network packets to
5946 start a database session.
5952 Improve trigger/constraint performance (Stephan)
5958 Improve speed of <literal>col IN (const, const, const, ...)</literal> (Tom)
5964 Fix hash indexes which were broken in rare cases (Tom)
5968 <listitem><para>Improve hash index concurrency and speed (Tom)</para>
5970 Prior releases suffered from poor hash index performance,
5971 particularly for high concurrency situations. This release fixes
5972 that, and the development group is interested in reports
5973 comparing B-tree and hash index performance.
5978 <para>Align shared buffers on 32-byte boundary for copy speed improvement (Manfred Spraul)</para>
5980 Certain CPU's perform faster data copies when addresses are
5986 <para>Data type <type>numeric</type> reimplemented for better performance (Tom)</para>
5988 <type>numeric</type> used to be stored in base 100. The new code
5989 uses base 10000, for significantly better performance.
5996 <title>Server Configuration Changes</title>
6000 <para>Rename server parameter <varname>server_min_messages</> to <varname>log_min_messages</> (Bruce)</para>
6002 This was done so most parameters that control the server logs
6003 begin with <literal>log_</>.
6007 <listitem><para>Rename <varname>show_*_stats</> to <varname>log_*_stats</> (Bruce)</para></listitem>
6008 <listitem><para>Rename <varname>show_source_port</> to <varname>log_source_port</> (Bruce)</para></listitem>
6009 <listitem><para>Rename <varname>hostname_lookup</> to <varname>log_hostname</> (Bruce)</para></listitem>
6012 <para>Add <varname>checkpoint_warning</> to warn of excessive checkpointing (Bruce)</para>
6014 In prior releases, it was difficult to determine if checkpoint
6015 was happening too frequently. This feature adds a warning to the
6016 server logs when excessive checkpointing happens.
6020 <listitem><para>New read-only server parameters for localization (Tom)</para></listitem>
6024 Change debug server log messages to output as <literal>DEBUG</>
6025 rather than <literal>LOG</> (Bruce)
6030 <para>Prevent server log variables from being turned off by non-superusers (Bruce)</para>
6032 This is a security feature so non-superusers cannot disable
6033 logging that was enabled by the administrator.
6039 <varname>log_min_messages</>/<varname>client_min_messages</> now
6040 controls <varname>debug_*</> output (Bruce)
6043 This centralizes client debug information so all debug output
6044 can be sent to either the client or server logs.
6049 <para>Add Mac OS X Rendezvous server support (Chris Campbell)</para>
6051 This allows Mac OS X hosts to query the network for available
6052 <productname>PostgreSQL</productname> servers.
6058 Add ability to print only slow statements using
6059 <varname>log_min_duration_statement</varname>
6063 This is an often requested debugging feature that allows
6064 administrators to see only slow queries in their server logs.
6069 <para>Allow <filename>pg_hba.conf</filename> to accept netmasks in CIDR format (Andrew Dunstan)</para>
6071 This allows administrators to merge the host IP address and
6072 netmask fields into a single CIDR field in <filename>pg_hba.conf</filename>.
6076 <listitem><para>New read-only parameter <varname>is_superuser</varname> (Tom)</para></listitem>
6079 <para>New parameter <varname>log_error_verbosity</varname> to control error detail (Tom)</para>
6081 This works with the new error reporting feature to supply
6082 additional error information like hints, file names and line
6088 <para><literal>postgres --describe-config</literal> now dumps server config variables (Aizaz Ahmed, Peter)</para>
6090 This option is useful for administration tools that need to know
6091 the configuration variable names and their minimums, maximums,
6092 defaults, and descriptions.
6098 Add new columns in <literal>pg_settings</literal>:
6099 <literal>context</>, <literal>type</>, <literal>source</>,
6100 <literal>min_val</>, <literal>max_val</> (Joe)
6106 Make default <varname>shared_buffers</> 1000 and
6107 <varname>max_connections</> 100, if possible (Tom)
6110 Prior versions defaulted to 64 shared buffers so <productname>PostgreSQL</productname>
6111 would start on even very old systems. This release tests the
6112 amount of shared memory allowed by the platform and selects more
6113 reasonable default values if possible. Of course, users are
6114 still encouraged to evaluate their resource load and size
6115 <varname>shared_buffers</varname> accordingly.
6121 New <filename>pg_hba.conf</filename> record type
6122 <literal>hostnossl</> to prevent SSL connections (Jon
6126 In prior releases, there was no way to prevent SSL connections
6127 if both the client and server supported SSL. This option allows
6134 Remove parameter <varname>geqo_random_seed</varname>
6141 Add server parameter <varname>regex_flavor</varname> to control regular expression processing (Tom)
6147 Make <command>pg_ctl</command> better handle nonstandard ports (Greg)
6154 <title>Query Changes</title>
6157 <listitem><para>New SQL-standard information schema (Peter)</para></listitem>
6158 <listitem><para>Add read-only transactions (Peter)</para></listitem>
6159 <listitem><para>Print key name and value in foreign-key violation messages (Dmitry Tkach)</para></listitem>
6162 <para>Allow users to see their own queries in <literal>pg_stat_activity</literal> (Kevin Brown)</para>
6164 In prior releases, only the superuser could see query strings
6165 using <literal>pg_stat_activity</literal>. Now ordinary users
6166 can see their own query strings.
6171 <para>Fix aggregates in subqueries to match SQL standard (Tom)</para>
6173 The SQL standard says that an aggregate function appearing
6174 within a nested subquery belongs to the outer query if its
6175 argument contains only outer-query variables. Prior
6176 <productname>PostgreSQL</productname> releases did not handle
6177 this fine point correctly.
6182 <para>Add option to prevent auto-addition of tables referenced in query (Nigel J. Andrews)</para>
6184 By default, tables mentioned in the query are automatically
6185 added to the <literal>FROM</> clause if they are not already
6186 there. This is compatible with historic
6187 <productname>POSTGRES</productname> behavior but is contrary to
6188 the SQL standard. This option allows selecting
6189 standard-compatible behavior.
6194 <para>Allow <literal>UPDATE ... SET col = DEFAULT</literal> (Rod)</para>
6196 This allows <command>UPDATE</command> to set a column to its
6197 declared default value.
6202 <para>Allow expressions to be used in <literal>LIMIT</>/<literal>OFFSET</> (Tom)</para>
6204 In prior releases, <literal>LIMIT</>/<literal>OFFSET</> could
6205 only use constants, not expressions.
6210 <para>Implement <literal>CREATE TABLE AS EXECUTE</literal> (Neil, Peter)</para>
6216 <title>Object Manipulation Changes</title>
6220 <para>Make <command>CREATE SEQUENCE</command> grammar more conforming to SQL:2003 (Neil)</para>
6224 <para>Add statement-level triggers (Neil)</para>
6226 While this allows a trigger to fire at the end of a statement,
6227 it does not allow the trigger to access all rows modified by the
6228 statement. This capability is planned for a future release.
6233 <para>Add check constraints for domains (Rod)</para>
6235 This greatly increases the usefulness of domains by allowing
6236 them to use check constraints.
6241 <para>Add <command>ALTER DOMAIN</command> (Rod)</para>
6243 This allows manipulation of existing domains.
6248 <para>Fix several zero-column table bugs (Tom)</para>
6250 <productname>PostgreSQL</productname> supports zero-column tables. This fixes various bugs
6251 that occur when using such tables.
6256 <para>Have <literal>ALTER TABLE ... ADD PRIMARY KEY</literal> add not-null constraint (Rod)</para>
6258 In prior releases, <literal>ALTER TABLE ... ADD
6259 PRIMARY</literal> would add a unique index, but not a not-null
6260 constraint. That is fixed in this release.
6264 <listitem><para>Add <literal>ALTER TABLE ... WITHOUT OIDS</literal> (Rod)</para>
6266 This allows control over whether new and updated rows will have
6267 an OID column. This is most useful for saving storage space.
6273 Add <literal>ALTER SEQUENCE</literal> to modify minimum, maximum,
6274 increment, cache, cycle values (Rod)
6279 <para>Add <literal>ALTER TABLE ... CLUSTER ON</literal> (Alvaro Herrera)</para>
6281 This command is used by <command>pg_dump</command> to record the
6282 cluster column for each table previously clustered. This
6283 information is used by database-wide cluster to cluster all
6284 previously clustered tables.
6288 <listitem><para>Improve automatic type casting for domains (Rod, Tom)</para></listitem>
6289 <listitem><para>Allow dollar signs in identifiers, except as first character (Tom)</para></listitem>
6290 <listitem><para>Disallow dollar signs in operator names, so <literal>x=$1</> works (Tom)</para></listitem>
6294 Allow copying table schema using <literal>LIKE
6295 <replaceable>subtable</replaceable></literal>, also SQL:2003
6296 feature <literal>INCLUDING DEFAULTS</literal> (Rod)
6302 Add <literal>WITH GRANT OPTION</literal> clause to
6303 <command>GRANT</command> (Peter)
6306 This enabled <command>GRANT</command> to give other users the
6307 ability to grant privileges on a object.
6314 <title>Utility Command Changes</title>
6318 <para>Add <literal>ON COMMIT</literal> clause to <command>CREATE TABLE</command> for temporary tables (Gavin)</para>
6320 This adds the ability for a table to be dropped or all rows
6321 deleted on transaction commit.
6326 <para>Allow cursors outside transactions using <literal>WITH HOLD</literal> (Neil)</para>
6328 In previous releases, cursors were removed at the end of the
6329 transaction that created them. Cursors can now be created with
6330 the <literal>WITH HOLD</literal> option, which allows them to
6331 continue to be accessed after the creating transaction has
6337 <para><literal>FETCH 0</literal> and <literal>MOVE 0 </literal> now do nothing (Bruce)</para>
6339 In previous releases, <literal>FETCH 0</literal> fetched all
6340 remaining rows, and <literal>MOVE 0</literal> moved to the end
6347 Cause <command>FETCH</command> and <command>MOVE</command> to
6348 return the number of rows fetched/moved, or zero if at the
6349 beginning/end of cursor, per SQL standard (Bruce)
6352 In prior releases, the row count returned by
6353 <command>FETCH</command> and <command>MOVE</command> did not
6354 accurately reflect the number of rows processed.
6359 <para>Properly handle <literal>SCROLL</literal> with cursors, or
6360 report an error (Neil)</para>
6362 Allowing random access (both forward and backward scrolling) to
6363 some kinds of queries cannot be done without some additional
6364 work. If <literal>SCROLL</literal> is specified when the cursor
6365 is created, this additional work will be performed. Furthermore,
6366 if the cursor has been created with <literal>NO SCROLL</literal>,
6367 no random access is allowed.
6373 Implement SQL-compatible options <literal>FIRST</>,
6374 <literal>LAST</>, <literal>ABSOLUTE <replaceable>n</></>,
6375 <literal>RELATIVE <replaceable>n</></> for
6376 <command>FETCH</command> and <command>MOVE</command> (Tom)
6381 <para>Allow <command>EXPLAIN</command> on <command>DECLARE CURSOR</command> (Tom)</para>
6385 <para>Allow <command>CLUSTER</command> to use index marked as pre-clustered by default (Alvaro Herrera)</para>
6389 <para>Allow <command>CLUSTER</command> to cluster all tables (Alvaro Herrera)</para>
6391 This allows all previously clustered tables in a database to be
6392 reclustered with a single command.
6396 <listitem><para>Prevent <command>CLUSTER</command> on partial indexes (Tom)</para></listitem>
6398 <listitem><para>Allow DOS and Mac line-endings in <command>COPY</> files (Bruce)</para></listitem>
6402 Disallow literal carriage return as a data value,
6403 backslash-carriage-return and <literal>\r</> are still allowed
6409 <para><command>COPY</> changes (binary, <literal>\.</>) (Tom)</para>
6413 <para>Recover from <command>COPY</command> failure cleanly (Tom)</para>
6417 <para>Prevent possible memory leaks in <command>COPY</command> (Tom)</para>
6421 <para>Make <command>TRUNCATE</command> transaction-safe (Rod)</para>
6423 <command>TRUNCATE</command> can now be used inside a
6424 transaction. If the transaction aborts, the changes made by the
6425 <command>TRUNCATE</command> are automatically rolled back.
6431 Allow prepare/bind of utility commands like
6432 <command>FETCH</command> and <command>EXPLAIN</command> (Tom)
6437 <para>Add <command>EXPLAIN EXECUTE</command> (Neil)</para>
6441 <para>Improve <command>VACUUM</command> performance on indexes by reducing WAL traffic (Tom)</para>
6445 <para>Functional indexes have been generalized into indexes on expressions (Tom)</para>
6447 In prior releases, functional indexes only supported a simple
6448 function applied to one or more column names. This release
6449 allows any type of scalar expression.
6455 Have <command>SHOW TRANSACTION ISOLATION</command> match input
6456 to <command>SET TRANSACTION ISOLATION</command>
6463 Have <command>COMMENT ON DATABASE</command> on nonlocal
6464 database generate a warning, rather than an error (Rod)
6468 Database comments are stored in database-local tables so
6469 comments on a database have to be stored in each database.
6475 Improve reliability of <command>LISTEN</>/<command>NOTIFY</> (Tom)
6480 <para>Allow <command>REINDEX</command> to reliably reindex nonshared system catalog indexes (Tom)</para>
6482 This allows system tables to be reindexed without the
6483 requirement of a standalone session, which was necessary in
6484 previous releases. The only tables that now require a standalone
6485 session for reindexing are the global system tables
6486 <literal>pg_database</>, <literal>pg_shadow</>, and
6487 <literal>pg_group</>.
6494 <title>Data Type and Function Changes</title>
6499 New server parameter <varname>extra_float_digits</varname> to
6500 control precision display of floating-point numbers (Pedro
6504 This controls output precision which was causing regression
6509 <listitem><para>Allow <literal>+1300</literal> as a numeric time-zone specifier, for FJST (Tom)</para></listitem>
6513 Remove rarely used functions <function>oidrand</>,
6514 <function>oidsrand</>, and <function>userfntest</> functions
6520 <para>Add <function>md5()</> function to main server, already in <filename>contrib/pgcrypto</filename> (Joe)</para>
6522 An MD5 function was frequently requested. For more complex
6523 encryption capabilities, use
6524 <filename>contrib/pgcrypto</filename>.
6528 <listitem><para>Increase date range of <type>timestamp</type> (John Cochran)</para></listitem>
6532 Change <literal>EXTRACT(EPOCH FROM timestamp)</literal> so
6533 <type>timestamp without time zone</type> is assumed to be in
6534 local time, not GMT (Tom)
6538 <listitem><para>Trap division by zero in case the operating system doesn't prevent it (Tom)</para></listitem>
6539 <listitem><para>Change the <type>numeric</type> data type internally to base 10000 (Tom)</para></listitem>
6540 <listitem><para>New <function>hostmask()</function> function (Greg Wickham)</para></listitem>
6541 <listitem><para>Fixes for <function>to_char()</function> and <function>to_timestamp()</function> (Karel)</para></listitem>
6545 Allow functions that can take any argument data type and return
6546 any data type, using <type>anyelement</type> and
6547 <type>anyarray</type> (Joe)
6550 This allows the creation of functions that can work with any
6557 Arrays may now be specified as <literal>ARRAY[1,2,3]</literal>,
6558 <literal>ARRAY[['a','b'],['c','d']]</literal>, or
6559 <literal>ARRAY[ARRAY[ARRAY[2]]]</literal> (Joe)
6565 Allow proper comparisons for arrays, including <literal>ORDER
6566 BY</literal> and <literal>DISTINCT</literal> support
6571 <listitem><para>Allow indexes on array columns (Joe)</para></listitem>
6572 <listitem><para>Allow array concatenation with <literal>||</literal> (Joe)</para></listitem>
6576 Allow <literal>WHERE</literal> qualification
6577 <literal><replaceable>expr</> <replaceable>op</> ANY/SOME/ALL
6578 (<replaceable>array_expr</>)</literal> (Joe)
6581 This allows arrays to behave like a list of values, for purposes
6582 like <literal>SELECT * FROM tab WHERE col IN
6583 (array_val)</literal>.
6589 New array functions <function>array_append</>,
6590 <function>array_cat</>, <function>array_lower</>,
6591 <function>array_prepend</>, <function>array_to_string</>,
6592 <function>array_upper</>, <function>string_to_array</> (Joe)
6596 <listitem><para>Allow user defined aggregates to use polymorphic functions (Joe)</para></listitem>
6597 <listitem><para>Allow assignments to empty arrays (Joe)</para></listitem>
6601 Allow 60 in seconds fields of <type>time</type>,
6602 <type>timestamp</type>, and <type>interval</type> input values
6606 Sixty-second values are needed for leap seconds.
6610 <listitem><para>Allow <type>cidr</type> data type to be cast to <type>text</type> (Tom)</para></listitem>
6612 <listitem><para>Disallow invalid time zone names in SET TIMEZONE</para></listitem>
6616 Trim trailing spaces when <type>char</type> is cast to
6617 <type>varchar</> or <type>text</> (Tom)
6623 Make <type>float(<replaceable>p</>)</> measure the precision
6624 <replaceable>p</> in binary digits, not decimal digits
6630 <para>Add IPv6 support to the <type>inet</type> and <type>cidr</type> data types (Michael Graff)</para>
6634 <para>Add <function>family()</function> function to report whether address is IPv4 or IPv6 (Michael Graff)</para>
6639 Have <literal>SHOW datestyle</literal> generate output similar
6640 to that used by <literal>SET datestyle</literal> (Tom)
6646 Make <literal>EXTRACT(TIMEZONE)</literal> and <literal>SET/SHOW
6647 TIME ZONE</literal> follow the SQL convention for the sign of
6648 time zone offsets, i.e., positive is east from UTC (Tom)
6653 <para>Fix <literal>date_trunc('quarter', ...)</literal> (Böjthe Zoltán)</para>
6655 Prior releases returned an incorrect value for this function call.
6660 <para>Make <function>initcap()</function> more compatible with Oracle (Mike Nolan)</para>
6662 <function>initcap()</function> now uppercases a letter appearing
6663 after any non-alphanumeric character, rather than only after
6669 <para>Allow only <varname>datestyle</varname> field order for date values not in ISO-8601 format (Greg)</para>
6674 Add new <varname>datestyle</varname> values <literal>MDY</>,
6675 <literal>DMY</>, and <literal>YMD</> to set input field order;
6676 honor <literal>US</> and <literal>European</> for backward
6683 String literals like <literal>'now'</literal> or
6684 <literal>'today'</literal> will no longer work as a column
6685 default. Use functions such as <function>now()</function>,
6686 <function>current_timestamp</function> instead. (change
6687 required for prepared statements) (Tom)
6692 <para>Treat NaN as larger than any other value in <function>min()</>/<function>max()</> (Tom)</para>
6694 NaN was already sorted after ordinary numeric values for most
6695 purposes, but <function>min()</> and <function>max()</> didn't
6701 <para>Prevent interval from suppressing <literal>:00</literal>
6702 seconds display</para>
6707 New functions <function>pg_get_triggerdef(prettyprint)</function>
6708 and <function>pg_conversion_is_visible()</function> (Christopher)
6713 <para>Allow time to be specified as <literal>040506</> or <literal>0405</> (Tom)</para>
6718 Input date order must now be <literal>YYYY-MM-DD</literal> (with 4-digit year) or
6719 match <varname>datestyle</varname>
6725 Make <function>pg_get_constraintdef</function> support
6726 unique, primary-key, and check constraints (Christopher)
6733 <title>Server-Side Language Changes</title>
6738 Prevent PL/pgSQL crash when <literal>RETURN NEXT</literal> is
6739 used on a zero-row record variable (Tom)
6745 Make PL/Python's <function>spi_execute</function> interface
6746 handle null values properly (Andrew Bosma)
6751 <para>Allow PL/pgSQL to declare variables of composite types without <literal>%ROWTYPE</literal> (Tom)</para>
6755 <para>Fix PL/Python's <function>_quote()</function> function to handle big integers</para>
6759 <para>Make PL/Python an untrusted language, now called <literal>plpythonu</literal> (Kevin Jacobs, Tom)</para>
6761 The Python language no longer supports a restricted execution
6762 environment, so the trusted version of PL/Python was removed. If
6763 this situation changes, a version of PL/Python that can be used
6764 by non-superusers will be readded.
6769 <para>Allow polymorphic PL/pgSQL functions (Joe, Tom)</para>
6773 <para>Allow polymorphic SQL functions (Joe)</para>
6778 Improved compiled function caching mechanism in PL/pgSQL with
6779 full support for polymorphism (Joe)
6785 Add new parameter <literal>$0</> in PL/pgSQL representing the
6786 function's actual return type (Joe)
6792 Allow PL/Tcl and PL/Python to use the same trigger on multiple tables (Tom)
6798 Fixed PL/Tcl's <function>spi_prepare</function> to accept fully
6799 qualified type names in the parameter type list
6807 <title>psql Changes</title>
6811 <para>Add <literal>\pset pager always</literal> to always use pager (Greg)</para>
6813 This forces the pager to be used even if the number of rows is
6814 less than the screen height. This is valuable for rows that
6815 wrap across several screen rows.
6819 <listitem><para>Improve tab completion (Rod, Ross Reedstrom, Ian Barwick)</para></listitem>
6820 <listitem><para>Reorder <literal>\?</> help into groupings (Harald Armin Massa, Bruce)</para></listitem>
6821 <listitem><para>Add backslash commands for listing schemas, casts, and conversions (Christopher)</para></listitem>
6825 <command>\encoding</> now changes based on the server parameter
6826 <varname>client_encoding</varname> (Tom)
6829 In previous versions, <command>\encoding</command> was not aware
6830 of encoding changes made using <literal>SET
6831 client_encoding</literal>.
6836 <para>Save editor buffer into readline history (Ross)</para>
6838 When <command>\e</> is used to edit a query, the result is saved
6839 in the readline history for retrieval using the up arrow.
6843 <listitem><para>Improve <command>\d</command> display (Christopher)</para></listitem>
6844 <listitem><para>Enhance HTML mode to be more standards-conforming (Greg)</para></listitem>
6847 <para>New <command>\set AUTOCOMMIT off</command> capability (Tom)</para>
6849 This takes the place of the removed server parameter <varname>autocommit</varname>.
6854 <para>New <command>\set VERBOSITY</command> to control error detail (Tom)</para>
6856 This controls the new error reporting details.
6860 <listitem><para>New prompt escape sequence <literal>%x</literal> to show transaction status (Tom)</para></listitem>
6861 <listitem><para>Long options for <application>psql</application> are now available on all platforms</para></listitem>
6866 <title>pg_dump Changes</title>
6869 <listitem><para>Multiple pg_dump fixes, including tar format and large objects</para></listitem>
6870 <listitem><para>Allow pg_dump to dump specific schemas (Neil)</para></listitem>
6873 <para>Make pg_dump preserve column storage characteristics (Christopher)</para>
6875 This preserves <literal>ALTER TABLE ... SET STORAGE</literal> information.
6879 <listitem><para>Make pg_dump preserve <command>CLUSTER</command> characteristics (Christopher)</para></listitem>
6883 Have pg_dumpall use <command>GRANT</>/<command>REVOKE</> to dump database-level privleges (Tom)
6889 Allow pg_dumpall to support the options <option>-a</>,
6890 <option>-s</>, <option>-x</> of pg_dump (Tom)
6894 <listitem><para>Prevent pg_dump from lowercasing identifiers specified on the command line (Tom)</para></listitem>
6898 pg_dump options <option>--use-set-session-authorization</option>
6899 and <option>--no-reconnect</option> now do nothing, all dumps
6900 use <command>SET SESSION AUTHORIZATION</command>
6903 pg_dump no longer reconnects to switch users, but instead always
6904 uses <command>SET SESSION AUTHORIZATION</command>. This will
6905 reduce password prompting during restores.
6910 <para>Long options for <application>pg_dump</application> are now available on all platforms</para>
6912 <productname>PostgreSQL</productname> now includes its own
6913 long-option processing routines.
6920 <title>libpq Changes</title>
6925 Add function <function>PQfreemem</function> for freeing memory on
6926 Windows, suggested for <command>NOTIFY</command> (Bruce)
6929 Windows requires that memory allocated in a library be freed by
6930 a function in the same library, hence
6931 <function>free()</function> doesn't work for freeing memory
6932 allocated by libpq. <function>PQfreemem</function> is the proper
6933 way to free libpq memory, especially on Windows, and is
6934 recommended for other platforms as well.
6939 <para>Document service capability, and add sample file (Bruce)</para>
6941 This allows clients to look up connection information in a
6942 central file on the client machine.
6948 Make <function>PQsetdbLogin</function> have the same defaults as
6949 <function>PQconnectdb</function> (Tom)
6953 <listitem><para>Allow libpq to cleanly fail when result sets are too large (Tom)</para></listitem>
6957 Improve performance of function <function>PQunescapeBytea</function> (Ben Lamb)
6963 Allow thread-safe libpq with <filename>configure</filename>
6964 option <option>--enable-thread-safety</option> (Lee Kindness,
6971 Allow function <function>pqInternalNotice</function> to accept a
6972 format string and arguments instead of just a preformatted
6973 message (Tom, Sean Chittenden)
6979 Control SSL negotiation with <literal>sslmode</literal> values
6980 <literal>disable</literal>, <literal>allow</literal>,
6981 <literal>prefer</literal>, and <literal>require</literal> (Jon
6987 <para>Allow new error codes and levels of text (Tom)</para>
6991 <para>Allow access to the underlying table and column of a query result (Tom)</para>
6993 This is helpful for query-builder applications that want to know
6994 the underlying table and column names associated with a specific
6999 <listitem><para>Allow access to the current transaction status (Tom)</para></listitem>
7000 <listitem><para>Add ability to pass binary data directly to the server (Tom)</para></listitem>
7004 Add function <function>PQexecPrepared</function> and
7005 <function>PQsendQueryPrepared</function> functions which perform
7006 bind/execute of previously prepared statements (Tom)
7013 <title>JDBC Changes</title>
7016 <listitem><para>Allow <function>setNull</function> on updateable result sets</para></listitem>
7017 <listitem><para>Allow <function>executeBatch</function> on a prepared statement (Barry)</para></listitem>
7018 <listitem><para>Support SSL connections (Barry)</para></listitem>
7019 <listitem><para>Handle schema names in result sets (Paul Sorenson)</para></listitem>
7020 <listitem><para>Add refcursor support (Nic Ferrier)</para></listitem>
7025 <title>Miscellaneous Interface Changes</title>
7029 <para>Prevent possible memory leak or core dump during libpgtcl shutdown (Tom)</para>
7032 <para>Add Informix compatibility to ECPG (Michael)</para>
7034 This allows ECPG to process embedded C programs that were
7035 written using certain Informix extensions.
7040 <para>Add type <type>decimal</type> to ECPG that is fixed length, for Informix (Michael)</para>
7045 Allow thread-safe embedded SQL programs with
7046 <filename>configure</filename> option
7047 <option>--enable-thread-safety</option> (Lee Kindness, Bruce)
7050 This allows multiple threads to access the database at the same
7056 <para>Moved Python client PyGreSQL to <ulink url="http://www.pygresql.org"></ulink> (Marc)</para>
7062 <title>Source Code Changes</title>
7065 <listitem><para>Prevent need for separate platform geometry regression result files (Tom)</para></listitem>
7066 <listitem><para>Improved PPC locking primitive (Reinhard Max)</para></listitem>
7067 <listitem><para>New function <function>palloc0</function> to allocate and clear memory (Bruce)</para></listitem>
7068 <listitem><para>Fix locking code for s390x CPU (64-bit) (Tom)</para></listitem>
7069 <listitem><para>Allow OpenBSD to use local ident credentials (William Ahern)</para></listitem>
7070 <listitem><para>Make query plan trees read-only to executor (Tom)</para></listitem>
7071 <listitem><para>Add Darwin startup scripts (David Wheeler)</para></listitem>
7072 <listitem><para>Allow libpq to compile with Borland C++ compiler (Lester Godwin, Karl Waclawek)</para></listitem>
7073 <listitem><para>Use our own version of <function>getopt_long()</function> if needed (Peter)</para></listitem>
7074 <listitem><para>Convert administration scripts to C (Peter)</para></listitem>
7075 <listitem><para> Bison >= 1.85 is now required to build the <productname>PostgreSQL</> grammar, if building from CVS</para></listitem>
7076 <listitem><para>Merge documentation into one book (Peter)</para></listitem>
7077 <listitem><para>Add Windows compatibility functions (Bruce)</para></listitem>
7078 <listitem><para>Allow client interfaces to compile under MinGW (Bruce)</para></listitem>
7079 <listitem><para>New <function>ereport()</function> function for error reporting (Tom)</para></listitem>
7080 <listitem><para>Support Intel compiler on Linux (Peter)</para></listitem>
7081 <listitem><para>Improve Linux startup scripts (Slawomir Sudnik, Darko Prenosil)</para></listitem>
7082 <listitem><para>Add support for AMD Opteron and Itanium (Jeffrey W. Baker, Bruce)</para></listitem>
7084 <para>Remove <option>--enable-recode</option> option from <command>configure</command></para>
7086 This was no longer needed now that we have <command>CREATE CONVERSION</command>.
7090 <para>Generate a compile error if spinlock code is not found (Bruce)</para>
7092 Platforms without spinlock code will now fail to compile, rather
7093 than silently using semaphores. This failure can be disabled
7094 with a new <command>configure</command> option.
7101 <title>Contrib Changes</title>
7104 <listitem><para>Change dbmirror license to BSD</para></listitem>
7105 <listitem><para>Improve earthdistance (Bruno Wolff III)</para></listitem>
7106 <listitem><para>Portability improvements to pgcrypto (Marko Kreen)</para></listitem>
7107 <listitem><para>Prevent crash in xml (John Gray, Michael Richards)</para></listitem>
7108 <listitem><para>Update oracle</para></listitem>
7109 <listitem><para>Update mysql</para></listitem>
7110 <listitem><para>Update cube (Bruno Wolff III)</para></listitem>
7111 <listitem><para>Update earthdistance to use cube (Bruno Wolff III)</para></listitem>
7112 <listitem><para>Update btree_gist (Oleg)</para></listitem>
7113 <listitem><para>New tsearch2 full-text search module (Oleg, Teodor)</para></listitem>
7114 <listitem><para>Add hash-based crosstab function to tablefuncs (Joe)</para></listitem>
7115 <listitem><para>Add serial column to order <function>connectby()</> siblings in tablefuncs (Nabil Sayegh,Joe)</para></listitem>
7116 <listitem><para>Add named persistent connections to dblink (Shridhar Daithanka)</para></listitem>
7117 <listitem><para>New pg_autovacuum allows automatic <command>VACUUM</command> (Matthew T. O'Connor)</para></listitem>
7118 <listitem><para>Make pgbench honor environment variables <envar>PGHOST</>, <envar>PGPORT</>, <envar>PGUSER</> (Tatsuo)</para></listitem>
7119 <listitem><para>Improve intarray (Teodor Sigaev)</para></listitem>
7120 <listitem><para>Improve pgstattuple (Rod)</para></listitem>
7121 <listitem><para>Fix bug in <function>metaphone()</function> in fuzzystrmatch</para></listitem>
7122 <listitem><para>Improve adddepend (Rod)</para></listitem>
7123 <listitem><para>Update spi/timetravel (Böjthe Zoltán)</para></listitem>
7124 <listitem><para>Fix dbase <option>-s</> option and improve non-ASCII handling (Thomas Behr, Márcio Smiderle)</para></listitem>
7125 <listitem><para>Remove array module because features now included by default (Joe)</para></listitem>
7131 <sect1 id="release-7-3-10">
7132 <title>Release 7.3.10</title>
7135 <title>Release date</title>
7136 <simpara>2005-05-09</simpara>
7140 This release contains a variety of fixes from 7.3.9, including several
7141 security-related issues.
7145 <title>Migration to version 7.3.10</title>
7148 A dump/restore is not required for those running 7.3.X. However,
7149 it is one possible way of handling a significant security problem
7150 that has been found in the initial contents of 7.3.X system
7151 catalogs. A dump/initdb/reload sequence using 7.3.10's initdb will
7152 automatically correct this problem.
7156 The security problem is that the built-in character set encoding
7157 conversion functions can be invoked from SQL commands by unprivileged
7158 users, but the functions were not designed for such use and are not
7159 secure against malicious choices of arguments. The fix involves changing
7160 the declared parameter list of these functions so that they can no longer
7161 be invoked from SQL commands. (This does not affect their normal use
7162 by the encoding conversion machinery.)
7163 It is strongly recommended that all installations repair this error,
7164 either by initdb or by following the manual repair procedure given
7165 below. The error at least allows unprivileged database users to crash
7166 their server process, and may allow unprivileged users to gain the
7167 privileges of a database superuser.
7171 If you wish not to do an initdb, perform the following procedure instead.
7172 As the database superuser, do:
7176 UPDATE pg_proc SET proargtypes[3] = 'internal'::regtype
7177 WHERE pronamespace = 11 AND pronargs = 5
7178 AND proargtypes[2] = 'cstring'::regtype;
7179 -- The command should report having updated 90 rows;
7180 -- if not, rollback and investigate instead of committing!
7186 The above procedure must be carried out in <emphasis>each</> database
7187 of an installation, including <literal>template1</>, and ideally
7188 including <literal>template0</> as well. If you do not fix the
7189 template databases then any subsequently created databases will contain
7190 the same error. <literal>template1</> can be fixed in the same way
7191 as any other database, but fixing <literal>template0</> requires
7192 additional steps. First, from any database issue
7194 UPDATE pg_database SET datallowconn = true WHERE datname = 'template0';
7196 Next connect to <literal>template0</> and perform the above repair
7197 procedure. Finally, do
7199 -- re-freeze template0:
7201 -- and protect it against future alterations:
7202 UPDATE pg_database SET datallowconn = false WHERE datname = 'template0';
7208 <title>Changes</title>
7211 <listitem><para>Change encoding function signature to prevent
7212 misuse</para></listitem>
7213 <listitem><para>Repair ancient race condition that allowed a transaction to be
7214 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
7215 than for other purposes</para>
7216 <para>This is an extremely serious bug since it could lead to apparent
7217 data inconsistencies being briefly visible to applications.</para></listitem>
7218 <listitem><para>Repair race condition between relation extension and
7220 <para>This could theoretically have caused loss of a page's worth of
7221 freshly-inserted data, although the scenario seems of very low probability.
7222 There are no known cases of it having caused more than an Assert failure.
7224 <listitem><para>Fix comparisons of <type>TIME WITH TIME ZONE</> values</para>
7226 The comparison code was wrong in the case where the
7227 <literal>--enable-integer-datetimes</> configuration switch had been used.
7228 NOTE: if you have an index on a <type>TIME WITH TIME ZONE</> column,
7229 it will need to be <command>REINDEX</>ed after installing this update, because
7230 the fix corrects the sort order of column values.
7232 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
7233 <type>TIME WITH TIME ZONE</> values</para></listitem>
7234 <listitem><para>Fix mis-display of negative fractional seconds in
7235 <type>INTERVAL</> values</para>
7237 This error only occurred when the
7238 <literal>--enable-integer-datetimes</> configuration switch had been used.
7240 <listitem><para>Additional buffer overrun checks in plpgsql
7241 (Neil)</para></listitem>
7242 <listitem><para>Fix pg_dump to dump trigger names containing <literal>%</>
7243 correctly (Neil)</para></listitem>
7244 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
7245 month-related formats</para></listitem>
7246 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
7247 (Marko Kreen)</para></listitem>
7248 <listitem><para>Still more 64-bit fixes for
7249 <filename>contrib/intagg</></para></listitem>
7250 <listitem><para>Prevent incorrect optimization of functions returning
7251 <type>RECORD</></para></listitem>
7257 <sect1 id="release-7-3-9">
7258 <title>Release 7.3.9</title>
7261 <title>Release date</title>
7262 <simpara>2005-01-31</simpara>
7266 This release contains a variety of fixes from 7.3.8, including several
7267 security-related issues.
7271 <title>Migration to version 7.3.9</title>
7274 A dump/restore is not required for those running 7.3.X.
7279 <title>Changes</title>
7282 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
7284 On platforms that will automatically execute initialization functions of a
7285 shared library (this includes at least Windows and ELF-based Unixen),
7286 <command>LOAD</> can be used to make the server execute arbitrary code.
7287 Thanks to NGS Software for reporting this.</para></listitem>
7288 <listitem><para>Check that creator of an aggregate function has the right to
7289 execute the specified transition functions</para>
7291 This oversight made it possible to bypass denial of EXECUTE
7292 permission on a function.</para></listitem>
7293 <listitem><para>Fix security and 64-bit issues in
7294 contrib/intagg</para></listitem>
7295 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
7296 Jurka)</para></listitem>
7297 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
7298 many parameters (Neil)</para></listitem>
7299 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
7301 The result of the join was mistakenly supposed to be sorted the same as the
7302 left input. This could not only deliver mis-sorted output to the user, but
7303 in case of nested merge joins could give outright wrong answers.
7305 <listitem><para>Fix plperl for quote marks in tuple fields</para></listitem>
7306 <listitem><para>Fix display of negative intervals in SQL and GERMAN
7307 datestyles</para></listitem>
7313 <sect1 id="release-7-3-8">
7314 <title>Release 7.3.8</title>
7317 <title>Release date</title>
7318 <simpara>2004-10-22</simpara>
7322 This release contains a variety of fixes from 7.3.7.
7327 <title>Migration to version 7.3.8</title>
7330 A dump/restore is not required for those running 7.3.X.
7335 <title>Changes</title>
7338 <listitem><para>Repair possible failure to update hint bits on disk</para>
7340 Under rare circumstances this oversight could lead to
7341 <quote>could not access transaction status</> failures, which qualifies
7342 it as a potential-data-loss bug.
7344 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
7346 Very large left joins using a hash join plan could fail to output unmatched
7347 left-side rows given just the right data distribution.
7349 <listitem><para>Disallow running pg_ctl as root</para>
7351 This is to guard against any possible security issues.
7353 <listitem><para>Avoid using temp files in /tmp in make_oidjoins_check</para>
7355 This has been reported as a security issue, though it's hardly worthy of
7356 concern since there is no reason for non-developers to use this script anyway.
7363 <sect1 id="release-7-3-7">
7364 <title>Release 7.3.7</title>
7367 <title>Release date</title>
7368 <simpara>2004-08-16</simpara>
7372 This release contains one critical fix over 7.3.6, and some minor items.
7377 <title>Migration to version 7.3.7</title>
7380 A dump/restore is not required for those running 7.3.X.
7385 <title>Changes</title>
7388 <listitem><para>Prevent possible loss of committed transactions during crash</para>
7390 Due to insufficient interlocking between transaction commit and checkpointing,
7391 it was possible for transactions committed just before the most recent
7392 checkpoint to be lost, in whole or in part, following a database crash and
7393 restart. This is a serious bug that has existed
7394 since <productname>PostgreSQL</productname> 7.1.
7396 <listitem><para>Remove asymmetrical word processing in tsearch (Teodor)</para></listitem>
7397 <listitem><para>Properly schema-qualify function names when pg_dump'ing a CAST</para></listitem>
7403 <sect1 id="release-7-3-6">
7404 <title>Release 7.3.6</title>
7407 <title>Release date</title>
7408 <simpara>2004-03-02</simpara>
7412 This release contains a variety of fixes from 7.3.5.
7417 <title>Migration to version 7.3.6</title>
7420 A dump/restore is <emphasis>not</emphasis> required for those
7427 <title>Changes</title>
7430 <listitem><para>Revert erroneous changes in rule permissions checking</para>
7431 <para>A patch applied in 7.3.3 to fix a corner case in rule permissions checks
7432 turns out to have disabled rule-related permissions checks in many
7433 not-so-corner cases. This would for example allow users to insert into views
7434 they weren't supposed to have permission to insert into. We have therefore
7435 reverted the 7.3.3 patch. The original bug will be fixed in 8.0.
7437 <listitem><para>Repair incorrect order of operations in
7438 GetNewTransactionId()</para>
7440 This bug could result in failure under out-of-disk-space conditions, including
7441 inability to restart even after disk space is freed.
7443 <listitem><para>Ensure configure selects -fno-strict-aliasing even when
7444 an external value for CFLAGS is supplied</para>
7446 On some platforms, building with -fstrict-aliasing causes bugs.
7448 <listitem><para>Make pg_restore handle 64-bit off_t correctly</para>
7450 This bug prevented proper restoration from archive files exceeding 4Gb.
7452 <listitem><para>Make contrib/dblink not assume that local and remote type OIDs
7453 match (Joe)</para></listitem>
7454 <listitem><para>Quote connectby()'s start_with argument properly (Joe)</para></listitem>
7455 <listitem><para>Don't crash when a rowtype argument to a plpgsql function is
7456 NULL</para></listitem>
7457 <listitem><para>Avoid generating invalid character encoding sequences in
7458 corner cases when planning LIKE operations</para></listitem>
7459 <listitem><para>Ensure text_position() cannot scan past end of source string
7460 in multibyte cases (Korea PostgreSQL Users' Group)</para></listitem>
7461 <listitem><para>Fix index optimization and selectivity estimates for LIKE
7462 operations on bytea columns (Joe)</para></listitem>
7468 <sect1 id="release-7-3-5">
7469 <title>Release 7.3.5</title>
7472 <title>Release date</title>
7473 <simpara>2003-12-03</simpara>
7477 This has a variety of fixes from 7.3.4.
7482 <title>Migration to version 7.3.5</title>
7485 A dump/restore is <emphasis>not</emphasis> required for those
7491 <title>Changes</title>
7494 <listitem><para>Force zero_damaged_pages to be on during recovery from WAL</para></listitem>
7495 <listitem><para>Prevent some obscure cases of <quote>variable not in subplan target lists</quote></para></listitem>
7496 <listitem><para>Force stats processes to detach from shared memory, ensuring cleaner shutdown</para></listitem>
7497 <listitem><para>Make PQescapeBytea and byteaout consistent with each other (Joe)</para></listitem>
7498 <listitem><para>Added missing SPI_finish() calls to dblink's get_tuple_of_interest() (Joe)</para></listitem>
7499 <listitem><para>Fix for possible foreign key violation when rule rewrites INSERT (Jan)</para></listitem>
7500 <listitem><para>Support qualified type names in PL/Tcl's spi_prepare command (Jan)</para></listitem>
7501 <listitem><para>Make pg_dump handle a procedural language handler located in pg_catalog</para></listitem>
7502 <listitem><para>Make pg_dump handle cases where a custom opclass is in another schema</para></listitem>
7503 <listitem><para>Make pg_dump dump binary-compatible casts correctly (Jan)</para></listitem>
7504 <listitem><para>Fix insertion of expressions containing subqueries into rule bodies</para></listitem>
7505 <listitem><para>Fix incorrect argument processing in clusterdb script (Anand Ranganathan)</para></listitem>
7506 <listitem><para>Fix problems with dropped columns in plpython triggers</para></listitem>
7507 <listitem><para>Repair problems with to_char() reading past end of its input string (Karel)</para></listitem>
7508 <listitem><para>Fix GB18030 mapping errors (Tatsuo)</para></listitem>
7509 <listitem><para>Fix several problems with SSL error handling and asynchronous SSL I/O</para></listitem>
7510 <listitem><para>Remove ability to bind a list of values to a single parameter in JDBC
7511 (prevents possible SQL-injection attacks)</para></listitem>
7512 <listitem><para>Fix some errors in HAVE_INT64_TIMESTAMP code paths</para></listitem>
7513 <listitem><para>Fix corner case for btree search in parallel with first root page split</para></listitem>
7519 <sect1 id="release-7-3-4">
7520 <title>Release 7.3.4</title>
7523 <title>Release date</title>
7524 <simpara>2003-07-24</simpara>
7528 This has a variety of fixes from 7.3.3.
7533 <title>Migration to version 7.3.4</title>
7536 A dump/restore is <emphasis>not</emphasis> required for those
7542 <title>Changes</title>
7545 <listitem><para>Repair breakage in timestamp-to-date conversion for dates before 2000</para></listitem>
7546 <listitem><para>Prevent rare possibility of server startup failure (Tom)</para></listitem>
7547 <listitem><para>Fix bugs in interval-to-time conversion (Tom)</para></listitem>
7548 <listitem><para>Add constraint names in a few places in pg_dump (Rod)</para></listitem>
7549 <listitem><para>Improve performance of functions with many parameters (Tom)</para></listitem>
7550 <listitem><para>Fix to_ascii() buffer overruns (Tom)</para></listitem>
7551 <listitem><para>Prevent restore of database comments from throwing an error (Tom)</para></listitem>
7552 <listitem><para>Work around buggy strxfrm() present in some Solaris releases (Tom)</para></listitem>
7553 <listitem><para>Properly escape jdbc setObject() strings to improve security (Barry)</para></listitem>
7559 <sect1 id="release-7-3-3">
7560 <title>Release 7.3.3</title>
7563 <title>Release date</title>
7564 <simpara>2003-05-22</simpara>
7568 This release contains a variety of fixes for version 7.3.2.
7572 <title>Migration to version 7.3.3</title>
7575 A dump/restore is <emphasis>not</emphasis> required for those
7576 running version 7.3.*.
7581 <title>Changes</title>
7584 <listitem><para>Repair sometimes-incorrect computation of StartUpID after a crash</para></listitem>
7585 <listitem><para>Avoid slowness with lots of deferred triggers in one transaction (Stephan)</para></listitem>
7586 <listitem><para>Don't lock referenced row when <command>UPDATE</command> doesn't change foreign key's value (Jan)</para></listitem>
7587 <listitem><para>Use <command>-fPIC</command> not <command>-fpic</command> on Sparc (Tom Callaway)</para></listitem>
7588 <listitem><para>Repair lack of schema-awareness in contrib/reindexdb</para></listitem>
7589 <listitem><para>Fix contrib/intarray error for zero-element result array (Teodor)</para></listitem>
7590 <listitem><para>Ensure createuser script will exit on control-C (Oliver)</para></listitem>
7591 <listitem><para>Fix errors when the type of a dropped column has itself been dropped</para></listitem>
7592 <listitem><para><command>CHECKPOINT</command> does not cause database panic on failure in noncritical steps</para></listitem>
7593 <listitem><para>Accept 60 in seconds fields of timestamp, time, interval input values</para></listitem>
7594 <listitem><para>Issue notice, not error, if <type>TIMESTAMP</type>,
7595 <type> TIME</type>, or <type>INTERVAL</type> precision too large</para></listitem>
7596 <listitem><para>Fix <function>abstime-to-time</function> cast function (fix is
7597 not applied unless you <application>initdb</application>)</para></listitem>
7598 <listitem><para>Fix <application>pg_proc</application> entry for
7599 <type>timestampt_izone</type> (fix is not applied unless you
7600 <application>initdb</application>)</para></listitem>
7601 <listitem><para>Make <function>EXTRACT(EPOCH FROM timestamp without time zone)</function> treat input as local time</para></listitem>
7602 <listitem><para><command>'now'::timestamptz</command> gave wrong answer if timezone changed earlier in transaction</para></listitem>
7603 <listitem><para><envar>HAVE_INT64_TIMESTAMP</envar> code for time with timezone overwrote its input</para></listitem>
7604 <listitem><para>Accept <command>GLOBAL TEMP/TEMPORARY</command> as a
7605 synonym for <command>TEMPORARY</command></para></listitem>
7606 <listitem><para>Avoid improper schema-privilege-check failure in foreign-key triggers</para></listitem>
7607 <listitem><para>Fix bugs in foreign-key triggers for <command>SET DEFAULT</command> action</para></listitem>
7608 <listitem><para>Fix incorrect time-qual check in row fetch for
7609 <command>UPDATE</command> and <command>DELETE</command> triggers</para></listitem>
7610 <listitem><para>Foreign-key clauses were parsed but ignored in
7611 <command>ALTER TABLE ADD COLUMN</command></para></listitem>
7612 <listitem><para>Fix createlang script breakage for case where handler function already exists</para></listitem>
7613 <listitem><para>Fix misbehavior on zero-column tables in <application>pg_dump</application>, COPY, ANALYZE, other places</para></listitem>
7614 <listitem><para>Fix misbehavior of <function>func_error()</function> on type names containing '%'</para></listitem>
7615 <listitem><para>Fix misbehavior of <function>replace()</function> on strings containing '%'</para></listitem>
7616 <listitem><para>Regular-expression patterns containing certain multibyte characters failed</para></listitem>
7617 <listitem><para>Account correctly for <command>NULL</command>s in more cases in join size estimation</para></listitem>
7618 <listitem><para>Avoid conflict with system definition of <function>isblank()</function> function or macro</para></listitem>
7619 <listitem><para>Fix failure to convert large code point values in EUC_TW conversions (Tatsuo)</para></listitem>
7620 <listitem><para>Fix error recovery for <function>SSL_read</function>/<function>SSL_write</function> calls</para></listitem>
7621 <listitem><para>Don't do early constant-folding of type coercion expressions</para></listitem>
7622 <listitem><para>Validate page header fields immediately after reading in any page</para></listitem>
7623 <listitem><para>Repair incorrect check for ungrouped variables in unnamed joins</para></listitem>
7624 <listitem><para>Fix buffer overrun in <function>to_ascii</function> (Guido Notari)</para></listitem>
7625 <listitem><para>contrib/ltree fixes (Teodor)</para></listitem>
7626 <listitem><para>Fix core dump in deadlock detection on machines where char is unsigned</para></listitem>
7627 <listitem><para>Avoid running out of buffers in many-way indexscan (bug introduced in 7.3)</para></listitem>
7628 <listitem><para>Fix planner's selectivity estimation functions to handle domains properly</para></listitem>
7629 <listitem><para>Fix <application>dbmirror</application> memory-allocation bug (Steven Singer)</para></listitem>
7630 <listitem><para>Prevent infinite loop in <function>ln(numeric)</function> due to roundoff error</para></listitem>
7631 <listitem><para><command>GROUP BY</command> got confused if there were multiple equal GROUP BY items</para></listitem>
7632 <listitem><para>Fix bad plan when inherited <command>UPDATE</command>/<command>DELETE</command> references another inherited table</para></listitem>
7633 <listitem><para>Prevent clustering on incomplete (partial or non-NULL-storing) indexes</para></listitem>
7634 <listitem><para>Service shutdown request at proper time if it arrives while still starting up</para></listitem>
7635 <listitem><para>Fix left-links in temporary indexes (could make backwards scans miss entries)</para></listitem>
7636 <listitem><para>Fix incorrect handling of client_encoding setting in postgresql.conf (Tatsuo)</para></listitem>
7637 <listitem><para>Fix failure to respond to <command>pg_ctl stop -m fast</command> after Async_NotifyHandler runs</para></listitem>
7638 <listitem><para>Fix SPI for case where rule contains multiple statements of the same type</para></listitem>
7639 <listitem><para>Fix problem with checking for wrong type of access privilege in rule query</para></listitem>
7640 <listitem><para>Fix problem with <command>EXCEPT</command> in <command>CREATE RULE</command></para></listitem>
7641 <listitem><para>Prevent problem with dropping temp tables having serial columns</para></listitem>
7642 <listitem><para>Fix replace_vars_with_subplan_refs failure in complex views</para></listitem>
7643 <listitem><para>Fix regexp slowness in single-byte encodings (Tatsuo)</para></listitem>
7644 <listitem><para>Allow qualified type names in <command>CREATE CAST</command>
7645 and <command> DROP CAST</command></para></listitem>
7646 <listitem><para>Accept <function>SETOF type[]</function>, which formerly had to
7647 be written <function>SETOF _type</function></para></listitem>
7648 <listitem><para>Fix <application>pg_dump</application> core dump in some cases with procedural languages</para></listitem>
7649 <listitem><para>Force ISO datestyle in <application>pg_dump</application> output, for portability (Oliver)</para></listitem>
7650 <listitem><para><application>pg_dump</application> failed to handle error return
7651 from <function>lo_read</function> (Oleg Drokin)</para></listitem>
7652 <listitem><para><application>pg_dumpall</application> failed with groups having no members (Nick Eskelinen)</para></listitem>
7653 <listitem><para><application>pg_dumpall</application> failed to recognize --globals-only switch</para></listitem>
7654 <listitem><para>pg_restore failed to restore blobs if -X disable-triggers is specified</para></listitem>
7655 <listitem><para>Repair intrafunction memory leak in plpgsql</para></listitem>
7656 <listitem><para>pltcl's <command>elog</command> command dumped core if given wrong parameters (Ian Harding)</para></listitem>
7657 <listitem><para>plpython used wrong value of <envar>atttypmod</envar> (Brad McLean)</para></listitem>
7658 <listitem><para>Fix improper quoting of boolean values in Python interface (D'Arcy)</para></listitem>
7659 <listitem><para>Added <function>addDataType()</function> method to PGConnection interface for JDBC</para></listitem>
7660 <listitem><para>Fixed various problems with updateable ResultSets for JDBC (Shawn Green)</para></listitem>
7661 <listitem><para>Fixed various problems with DatabaseMetaData for JDBC (Kris Jurka, Peter Royal)</para></listitem>
7662 <listitem><para>Fixed problem with parsing table ACLs in JDBC</para></listitem>
7663 <listitem><para>Better error message for character set conversion problems in JDBC</para></listitem>
7669 <sect1 id="release-7-3-2">
7670 <title>Release 7.3.2</title>
7673 <title>Release date</title>
7674 <simpara>2003-02-04</simpara>
7678 This release contains a variety of fixes for version 7.3.1.
7683 <title>Migration to version 7.3.2</title>
7686 A dump/restore is <emphasis>not</emphasis> required for those
7687 running version 7.3.*.
7692 <title>Changes</title>
7695 <listitem><para>Restore creation of OID column in CREATE TABLE AS / SELECT INTO</para></listitem>
7696 <listitem><para>Fix <application>pg_dump</> core dump when dumping views having comments</para></listitem>
7697 <listitem><para>Dump DEFERRABLE/INITIALLY DEFERRED constraints properly</para></listitem>
7698 <listitem><para>Fix UPDATE when child table's column numbering differs from parent</para></listitem>
7699 <listitem><para>Increase default value of max_fsm_relations</para></listitem>
7700 <listitem><para>Fix problem when fetching backwards in a cursor for a single-row query</para></listitem>
7701 <listitem><para>Make backward fetch work properly with cursor on SELECT DISTINCT query</para></listitem>
7702 <listitem><para>Fix problems with loading <application>pg_dump</> files containing contrib/lo usage</para></listitem>
7703 <listitem><para>Fix problem with all-numeric user names</para></listitem>
7704 <listitem><para>Fix possible memory leak and core dump during disconnect in libpgtcl</para></listitem>
7705 <listitem><para>Make plpython's spi_execute command handle nulls properly (Andrew Bosma)</para></listitem>
7706 <listitem><para>Adjust plpython error reporting so that its regression test passes again</para></listitem>
7707 <listitem><para>Work with bison 1.875</para></listitem>
7708 <listitem><para>Handle mixed-case names properly in plpgsql's %type (Neil)</para></listitem>
7709 <listitem><para>Fix core dump in pltcl when executing a query rewritten by a rule</para></listitem>
7710 <listitem><para>Repair array subscript overruns (per report from Yichen Xie)</para></listitem>
7711 <listitem><para>Reduce MAX_TIME_PRECISION from 13 to 10 in floating-point case</para></listitem>
7712 <listitem><para>Correctly case-fold variable names in per-database and per-user settings</para></listitem>
7713 <listitem><para>Fix coredump in plpgsql's RETURN NEXT when SELECT into record returns no rows</para></listitem>
7714 <listitem><para>Fix outdated use of pg_type.typprtlen in python client interface</para></listitem>
7715 <listitem><para>Correctly handle fractional seconds in timestamps in JDBC driver</para></listitem>
7716 <listitem><para>Improve performance of getImportedKeys() in JDBC</para></listitem>
7717 <listitem><para>Make shared-library symlinks work standardly on HPUX (Giles)</para></listitem>
7718 <listitem><para>Repair inconsistent rounding behavior for timestamp, time, interval</para></listitem>
7719 <listitem><para>SSL negotiation fixes (Nathan Mueller)</para></listitem>
7720 <listitem><para>Make libpq's ~/.pgpass feature work when connecting with PQconnectDB</para></listitem>
7721 <listitem><para>Update my2pg, ora2pg</para></listitem>
7722 <listitem><para>Translation updates</para></listitem>
7723 <listitem><para>Add casts between types lo and oid in contrib/lo</para></listitem>
7724 <listitem><para>fastpath code now checks for privilege to call function</para></listitem>
7730 <sect1 id="release-7-3-1">
7731 <title>Release 7.3.1</title>
7734 <title>Release date</title>
7735 <simpara>2002-12-18</simpara>
7739 This release contains a variety of fixes for version 7.3.
7744 <title>Migration to version 7.3.1</title>
7747 A dump/restore is <emphasis>not</emphasis> required for those
7748 running version 7.3. However, it should be noted that the main
7749 <productname>PostgreSQL</productname> interface library, libpq,
7750 has a new major version number for this release, which may require
7751 recompilation of client code in certain cases.
7756 <title>Changes</title>
7759 <listitem><para>Fix a core dump of COPY TO when client/server encodings don't match (Tom)</para></listitem>
7760 <listitem><para>Allow <application>pg_dump</> to work with pre-7.2 servers (Philip)</para></listitem>
7761 <listitem><para>contrib/adddepend fixes (Tom)</para></listitem>
7762 <listitem><para>Fix problem with deletion of per-user/per-database config settings (Tom)</para></listitem>
7763 <listitem><para>contrib/vacuumlo fix (Tom)</para></listitem>
7764 <listitem><para>Allow 'password' encryption even when pg_shadow contains MD5 passwords (Bruce)</para></listitem>
7765 <listitem><para>contrib/dbmirror fix (Steven Singer)</para></listitem>
7766 <listitem><para>Optimizer fixes (Tom)</para></listitem>
7767 <listitem><para>contrib/tsearch fixes (Teodor Sigaev, Magnus)</para></listitem>
7768 <listitem><para>Allow locale names to be mixed case (Nicolai Tufar)</para></listitem>
7769 <listitem><para>Increment libpq library's major version number (Bruce)</para></listitem>
7770 <listitem><para>pg_hba.conf error reporting fixes (Bruce, Neil)</para></listitem>
7771 <listitem><para>Add SCO Openserver 5.0.4 as a supported platform (Bruce)</para></listitem>
7772 <listitem><para>Prevent EXPLAIN from crashing server (Tom)</para></listitem>
7773 <listitem><para>SSL fixes (Nathan Mueller)</para></listitem>
7774 <listitem><para>Prevent composite column creation via ALTER TABLE (Tom)</para></listitem>
7780 <sect1 id="release-7-3">
7781 <title>Release 7.3</title>
7784 <title>Release date</title>
7785 <simpara>2002-11-27</simpara>
7789 <title>Overview</title>
7792 Major changes in this release:
7797 <term>Schemas</term>
7800 Schemas allow users to create objects in separate namespaces,
7801 so two people or applications can have tables with the same
7802 name. There is also a public schema for shared tables.
7803 Table/index creation can be restricted by removing privileges
7804 on the public schema.
7810 <term>Drop Column</term>
7813 <productname>PostgreSQL</productname> now supports the
7814 <literal>ALTER TABLE ... DROP COLUMN</literal> functionality.
7820 <term>Table Functions</term>
7823 Functions returning multiple rows and/or multiple columns are
7824 now much easier to use than before. You can call such a
7825 <quote>table function</quote> in the <literal>SELECT</literal>
7826 <literal>FROM</literal> clause, treating its output like a
7827 table. Also, <application>PL/pgSQL</application> functions can
7834 <term>Prepared Queries</term>
7837 <productname>PostgreSQL</productname> now supports prepared
7838 queries, for improved performance.
7844 <term>Dependency Tracking</term>
7847 <productname>PostgreSQL</productname> now records object
7848 dependencies, which allows improvements in many areas.
7849 <command>DROP</command> statements now take either
7850 <literal>CASCADE</> or <literal>RESTRICT</> to control whether
7851 dependent objects are also dropped.
7857 <term>Privileges</term>
7860 Functions and procedural languages now have privileges, and
7861 functions can be defined to run with the privileges of their
7868 <term>Internationalization</term>
7871 Both multibyte and locale support are now always enabled.
7877 <term>Logging</term>
7880 A variety of logging options have been enhanced.
7886 <term>Interfaces</term>
7889 A large number of interfaces have been moved to <ulink
7890 url="http://gborg.postgresql.org">http://gborg.postgresql.org</>
7891 where they can be developed and released independently.
7897 <term>Functions/Identifiers</term>
7900 By default, functions can now take up to 32 parameters, and
7901 identifiers can be up to 63 bytes long. Also, <literal>OPAQUE</>
7902 is now deprecated: there are specific <quote>pseudo-datatypes</>
7903 to represent each of the former meanings of <literal>OPAQUE</>
7904 in function argument and result types.
7913 <title>Migration to version 7.3</title>
7916 A dump/restore using <application>pg_dump</> is required for those
7917 wishing to migrate data from any previous release. If your
7918 application examines the system catalogs, additional changes will
7919 be required due to the introduction of schemas in 7.3; for more
7920 information, see: <ulink
7921 url="http://developer.postgresql.org/~momjian/upgrade_tips_7.3"></>.
7925 Observe the following incompatibilities:
7931 Pre-6.3 clients are no longer supported.
7937 <filename>pg_hba.conf</filename> now has a column for the user
7938 name and additional features. Existing files need to be
7945 Several <filename>postgresql.conf</filename> logging parameters
7952 <literal>LIMIT #,#</literal> has been disabled; use
7953 <literal>LIMIT # OFFSET #</literal>.
7959 <command>INSERT</command> statements with column lists must
7960 specify a value for each specified column. For example,
7961 <literal>INSERT INTO tab (col1, col2) VALUES ('val1')</literal>
7962 is now invalid. It's still allowed to supply fewer columns than
7963 expected if the <command>INSERT</command> does not have a column list.
7969 <type>serial</type> columns are no longer automatically
7970 <literal>UNIQUE</>; thus, an index will not automatically be
7977 A <command>SET</command> command inside an aborted transaction
7984 <command>COPY</command> no longer considers missing trailing
7985 columns to be null. All columns need to be specified.
7986 (However, one may achieve a similar effect by specifying a
7987 column list in the <command>COPY</command> command.)
7993 The data type <type>timestamp</type> is now equivalent to
7994 <type>timestamp without time zone</type>, instead of
7995 <type>timestamp with time zone</type>.
8001 Pre-7.3 databases loaded into 7.3 will not have the new object
8002 dependencies for <type>serial</type> columns, unique
8003 constraints, and foreign keys. See the directory
8004 <filename>contrib/adddepend/</filename> for a detailed
8005 description and a script that will add such dependencies.
8011 An empty string (<literal>''</literal>) is no longer allowed as
8012 the input into an integer field. Formerly, it was silently
8021 <title>Changes</title>
8024 <title>Server Operation</title>
8026 <listitem><para>Add pg_locks view to show locks (Neil)</para></listitem>
8027 <listitem><para>Security fixes for password negotiation memory allocation (Neil)</para></listitem>
8028 <listitem><para>Remove support for version 0 FE/BE protocol (<productname>PostgreSQL</productname> 6.2 and earlier) (Tom)</para></listitem>
8029 <listitem><para>Reserve the last few backend slots for superusers, add parameter superuser_reserved_connections to control this (Nigel J. Andrews)</para></listitem>
8034 <title>Performance</title>
8036 <listitem><para>Improve startup by calling localtime() only once (Tom)</para></listitem>
8037 <listitem><para>Cache system catalog information in flat files for faster startup (Tom)</para></listitem>
8038 <listitem><para>Improve caching of index information (Tom)</para></listitem>
8039 <listitem><para>Optimizer improvements (Tom, Fernando Nasser)</para></listitem>
8040 <listitem><para>Catalog caches now store failed lookups (Tom)</para></listitem>
8041 <listitem><para>Hash function improvements (Neil)</para></listitem>
8042 <listitem><para>Improve performance of query tokenization and network handling (Peter)</para></listitem>
8043 <listitem><para>Speed improvement for large object restore (Mario Weilguni)</para></listitem>
8044 <listitem><para>Mark expired index entries on first lookup, saving later heap fetches (Tom)</para></listitem>
8045 <listitem><para>Avoid excessive NULL bitmap padding (Manfred Koizar)</para></listitem>
8046 <listitem><para>Add BSD-licensed qsort() for Solaris, for performance (Bruce)</para></listitem>
8047 <listitem><para>Reduce per-row overhead by four bytes (Manfred Koizar)</para></listitem>
8048 <listitem><para>Fix GEQO optimizer bug (Neil Conway)</para></listitem>
8049 <listitem><para>Make WITHOUT OID actually save four bytes per row (Manfred Koizar)</para></listitem>
8050 <listitem><para>Add default_statistics_target variable to specify ANALYZE buckets (Neil)</para></listitem>
8051 <listitem><para>Use local buffer cache for temporary tables so no WAL overhead (Tom)</para></listitem>
8052 <listitem><para>Improve free space map performance on large tables (Stephen Marshall, Tom)</para></listitem>
8053 <listitem><para>Improved WAL write concurrency (Tom)</para></listitem>
8058 <title>Privileges</title>
8060 <listitem><para>Add privileges on functions and procedural languages (Peter)</para></listitem>
8061 <listitem><para>Add OWNER to CREATE DATABASE so superusers can create databases on behalf of unprivileged users (Gavin Sherry, Tom)</para></listitem>
8062 <listitem><para>Add new object privilege bits EXECUTE and USAGE (Tom)</para></listitem>
8063 <listitem><para>Add SET SESSION AUTHORIZATION DEFAULT and RESET SESSION AUTHORIZATION (Tom)</para></listitem>
8064 <listitem><para>Allow functions to be executed with the privilege of the function owner (Peter)</para></listitem>
8069 <title>Server Configuration</title>
8071 <listitem><para>Server log messages now tagged with LOG, not DEBUG (Bruce)</para></listitem>
8072 <listitem><para>Add user column to pg_hba.conf (Bruce)</para></listitem>
8073 <listitem><para>Have log_connections output two lines in log file (Tom)</para></listitem>
8074 <listitem><para>Remove debug_level from postgresql.conf, now server_min_messages (Bruce)</para></listitem>
8075 <listitem><para>New ALTER DATABASE/USER ... SET command for per-user/database initialization (Peter)</para></listitem>
8076 <listitem><para>New parameters server_min_messages and client_min_messages to control which messages are sent to the server logs or client applications (Bruce)</para></listitem>
8077 <listitem><para>Allow pg_hba.conf to specify lists of users/databases separated by commas, group names prepended with +, and file names prepended with @ (Bruce)</para></listitem>
8078 <listitem><para>Remove secondary password file capability and pg_password utility (Bruce)</para></listitem>
8079 <listitem><para>Add variable db_user_namespace for database-local user names (Bruce)</para></listitem>
8080 <listitem><para>SSL improvements (Bear Giles)</para></listitem>
8081 <listitem><para>Make encryption of stored passwords the default (Bruce)</para></listitem>
8082 <listitem><para>Allow pg_statistics to be reset by calling pg_stat_reset() (Christopher)</para></listitem>
8083 <listitem><para>Add log_duration parameter (Bruce)</para></listitem>
8084 <listitem><para>Rename debug_print_query to log_statement (Bruce)</para></listitem>
8085 <listitem><para>Rename show_query_stats to show_statement_stats (Bruce)</para></listitem>
8086 <listitem><para>Add param log_min_error_statement to print commands to logs on error (Gavin)</para></listitem>
8091 <title>Queries</title>
8093 <listitem><para>Make cursors insensitive, meaning their contents do not change (Tom)</para></listitem>
8094 <listitem><para>Disable LIMIT #,# syntax; now only LIMIT # OFFSET # supported (Bruce)</para></listitem>
8095 <listitem><para>Increase identifier length to 63 (Neil, Bruce)</para></listitem>
8096 <listitem><para>UNION fixes for merging >= 3 columns of different lengths (Tom)</para></listitem>
8097 <listitem><para>Add DEFAULT key word to INSERT, e.g., INSERT ... (..., DEFAULT, ...) (Rod)</para></listitem>
8098 <listitem><para>Allow views to have default values using ALTER COLUMN ... SET DEFAULT (Neil)</para></listitem>
8099 <listitem><para>Fail on INSERTs with column lists that don't supply all column values, e.g., INSERT INTO tab (col1, col2) VALUES ('val1'); (Rod)</para></listitem>
8100 <listitem><para>Fix for join aliases (Tom)</para></listitem>
8101 <listitem><para>Fix for FULL OUTER JOINs (Tom)</para></listitem>
8102 <listitem><para>Improve reporting of invalid identifier and location (Tom, Gavin)</para></listitem>
8103 <listitem><para>Fix OPEN cursor(args) (Tom)</para></listitem>
8104 <listitem><para>Allow 'ctid' to be used in a view and currtid(viewname) (Hiroshi)</para></listitem>
8105 <listitem><para>Fix for CREATE TABLE AS with UNION (Tom)</para></listitem>
8106 <listitem><para>SQL99 syntax improvements (Thomas)</para></listitem>
8107 <listitem><para>Add statement_timeout variable to cancel queries (Bruce)</para></listitem>
8108 <listitem><para>Allow prepared queries with PREPARE/EXECUTE (Neil)</para></listitem>
8109 <listitem><para>Allow FOR UPDATE to appear after LIMIT/OFFSET (Bruce)</para></listitem>
8110 <listitem><para>Add variable autocommit (Tom, David Van Wie)</para></listitem>
8115 <title>Object Manipulation</title>
8117 <listitem><para>Make equals signs optional in CREATE DATABASE (Gavin Sherry)</para></listitem>
8118 <listitem><para>Make ALTER TABLE OWNER change index ownership too (Neil)</para></listitem>
8119 <listitem><para>New ALTER TABLE tabname ALTER COLUMN colname SET STORAGE controls TOAST storage, compression (John Gray)</para></listitem>
8120 <listitem><para>Add schema support, CREATE/DROP SCHEMA (Tom)</para></listitem>
8121 <listitem><para>Create schema for temporary tables (Tom)</para></listitem>
8122 <listitem><para>Add variable search_path for schema search (Tom)</para></listitem>
8123 <listitem><para>Add ALTER TABLE SET/DROP NOT NULL (Christopher)</para></listitem>
8124 <listitem><para>New CREATE FUNCTION volatility levels (Tom)</para></listitem>
8125 <listitem><para>Make rule names unique only per table (Tom)</para></listitem>
8126 <listitem><para>Add 'ON tablename' clause to DROP RULE and COMMENT ON RULE (Tom)</para></listitem>
8127 <listitem><para>Add ALTER TRIGGER RENAME (Joe)</para></listitem>
8128 <listitem><para>New current_schema() and current_schemas() inquiry functions (Tom)</para></listitem>
8129 <listitem><para>Allow functions to return multiple rows (table functions) (Joe)</para></listitem>
8130 <listitem><para>Make WITH optional in CREATE DATABASE, for consistency (Bruce)</para></listitem>
8131 <listitem><para>Add object dependency tracking (Rod, Tom)</para></listitem>
8132 <listitem><para>Add RESTRICT/CASCADE to DROP commands (Rod)</para></listitem>
8133 <listitem><para>Add ALTER TABLE DROP for non-CHECK CONSTRAINT (Rod)</para></listitem>
8134 <listitem><para>Autodestroy sequence on DROP of table with SERIAL (Rod)</para></listitem>
8135 <listitem><para>Prevent column dropping if column is used by foreign key (Rod)</para></listitem>
8136 <listitem><para>Automatically drop constraints/functions when object is dropped (Rod)</para></listitem>
8137 <listitem><para>Add CREATE/DROP OPERATOR CLASS (Bill Studenmund, Tom)</para></listitem>
8138 <listitem><para>Add ALTER TABLE DROP COLUMN (Christopher, Tom, Hiroshi)</para></listitem>
8139 <listitem><para>Prevent inherited columns from being removed or renamed (Alvaro Herrera)</para></listitem>
8140 <listitem><para>Fix foreign key constraints to not error on intermediate database states (Stephan)</para></listitem>
8141 <listitem><para>Propagate column or table renaming to foreign key constraints</para></listitem>
8142 <listitem><para>Add CREATE OR REPLACE VIEW (Gavin, Neil, Tom)</para></listitem>
8143 <listitem><para>Add CREATE OR REPLACE RULE (Gavin, Neil, Tom)</para></listitem>
8144 <listitem><para>Have rules execute alphabetically, returning more predictable values (Tom)</para></listitem>
8145 <listitem><para>Triggers are now fired in alphabetical order (Tom)</para></listitem>
8146 <listitem><para>Add /contrib/adddepend to handle pre-7.3 object dependencies (Rod)</para></listitem>
8147 <listitem><para>Allow better casting when inserting/updating values (Tom)</para></listitem>
8152 <title>Utility Commands</title>
8154 <listitem><para>Have COPY TO output embedded carriage returns and newlines as \r and \n (Tom)</para></listitem>
8155 <listitem><para>Allow DELIMITER in COPY FROM to be 8-bit clean (Tatsuo)</para></listitem>
8156 <listitem><para>Make <application>pg_dump</> use ALTER TABLE ADD PRIMARY KEY, for performance (Neil)</para></listitem>
8157 <listitem><para>Disable brackets in multistatement rules (Bruce)</para></listitem>
8158 <listitem><para>Disable VACUUM from being called inside a function (Bruce)</para></listitem>
8159 <listitem><para>Allow dropdb and other scripts to use identifiers with spaces (Bruce)</para></listitem>
8160 <listitem><para>Restrict database comment changes to the current database</para></listitem>
8161 <listitem><para>Allow comments on operators, independent of the underlying function (Rod)</para></listitem>
8162 <listitem><para>Rollback SET commands in aborted transactions (Tom)</para></listitem>
8163 <listitem><para>EXPLAIN now outputs as a query (Tom)</para></listitem>
8164 <listitem><para>Display condition expressions and sort keys in EXPLAIN (Tom)</para></listitem>
8165 <listitem><para>Add 'SET LOCAL var = value' to set configuration variables for a single transaction (Tom)</para></listitem>
8166 <listitem><para>Allow ANALYZE to run in a transaction (Bruce)</para></listitem>
8167 <listitem><para>Improve COPY syntax using new WITH clauses, keep backward compatibility (Bruce)</para></listitem>
8168 <listitem><para>Fix <application>pg_dump</> to consistently output tags in non-ASCII dumps (Bruce)</para></listitem>
8169 <listitem><para>Make foreign key constraints clearer in dump file (Rod)</para></listitem>
8170 <listitem><para>Add COMMENT ON CONSTRAINT (Rod)</para></listitem>
8171 <listitem><para>Allow COPY TO/FROM to specify column names (Brent Verner)</para></listitem>
8172 <listitem><para>Dump UNIQUE and PRIMARY KEY constraints as ALTER TABLE (Rod)</para></listitem>
8173 <listitem><para>Have SHOW output a query result (Joe)</para></listitem>
8174 <listitem><para>Generate failure on short COPY lines rather than pad NULLs (Neil)</para></listitem>
8175 <listitem><para>Fix CLUSTER to preserve all table attributes (Alvaro Herrera)</para></listitem>
8176 <listitem><para>New pg_settings table to view/modify GUC settings (Joe)</para></listitem>
8177 <listitem><para>Add smart quoting, portability improvements to <application>pg_dump</> output (Peter)</para></listitem>
8178 <listitem><para>Dump serial columns out as SERIAL (Tom)</para></listitem>
8179 <listitem><para>Enable large file support, >2G for <application>pg_dump</> (Peter, Philip Warner, Bruce)</para></listitem>
8180 <listitem><para>Disallow TRUNCATE on tables that are involved in referential constraints (Rod)</para></listitem>
8181 <listitem><para>Have TRUNCATE also auto-truncate the toast table of the relation (Tom)</para></listitem>
8182 <listitem><para>Add clusterdb utility that will auto-cluster an entire database based on previous CLUSTER operations (Alvaro Herrera)</para></listitem>
8183 <listitem><para>Overhaul pg_dumpall (Peter)</para></listitem>
8184 <listitem><para>Allow REINDEX of TOAST tables (Tom)</para></listitem>
8185 <listitem><para>Implemented START TRANSACTION, per SQL99 (Neil)</para></listitem>
8186 <listitem><para>Fix rare index corruption when a page split affects bulk delete (Tom)</para></listitem>
8187 <listitem><para>Fix ALTER TABLE ... ADD COLUMN for inheritance (Alvaro Herrera)</para></listitem>
8192 <title>Data Types and Functions</title>
8194 <listitem><para>Fix factorial(0) to return 1 (Bruce)</para></listitem>
8195 <listitem><para>Date/time/timezone improvements (Thomas)</para></listitem>
8196 <listitem><para>Fix for array slice extraction (Tom)</para></listitem>
8197 <listitem><para>Fix extract/date_part to report proper microseconds for timestamp (Tatsuo)</para></listitem>
8198 <listitem><para>Allow text_substr() and bytea_substr() to read TOAST values more efficiently (John Gray)</para></listitem>
8199 <listitem><para>Add domain support (Rod)</para></listitem>
8200 <listitem><para>Make WITHOUT TIME ZONE the default for TIMESTAMP and TIME data types (Thomas)</para></listitem>
8201 <listitem><para>Allow alternate storage scheme of 64-bit integers for date/time types using --enable-integer-datetimes in configure (Thomas)</para></listitem>
8202 <listitem><para>Make timezone(timestamptz) return timestamp rather than a string (Thomas)</para></listitem>
8203 <listitem><para>Allow fractional seconds in date/time types for dates prior to 1BC (Thomas)</para></listitem>
8204 <listitem><para>Limit timestamp data types to 6 decimal places of precision (Thomas)</para></listitem>
8205 <listitem><para>Change timezone conversion functions from timetz() to timezone() (Thomas)</para></listitem>
8206 <listitem><para>Add configuration variables datestyle and timezone (Tom)</para></listitem>
8207 <listitem><para>Add OVERLAY(), which allows substitution of a substring in a string (Thomas)</para></listitem>
8208 <listitem><para>Add SIMILAR TO (Thomas, Tom)</para></listitem>
8209 <listitem><para>Add regular expression SUBSTRING(string FROM pat FOR escape) (Thomas)</para></listitem>
8210 <listitem><para>Add LOCALTIME and LOCALTIMESTAMP functions (Thomas)</para></listitem>
8211 <listitem><para>Add named composite types using CREATE TYPE typename AS (column) (Joe)</para></listitem>
8212 <listitem><para>Allow composite type definition in the table alias clause (Joe)</para></listitem>
8213 <listitem><para>Add new API to simplify creation of C language table functions (Joe)</para></listitem>
8214 <listitem><para>Remove ODBC-compatible empty parentheses from calls to SQL99 functions for which these parentheses do not match the standard (Thomas)</para></listitem>
8215 <listitem><para>Allow macaddr data type to accept 12 hex digits with no separators (Mike Wyer)</para></listitem>
8216 <listitem><para>Add CREATE/DROP CAST (Peter)</para></listitem>
8217 <listitem><para>Add IS DISTINCT FROM operator (Thomas)</para></listitem>
8218 <listitem><para>Add SQL99 TREAT() function, synonym for CAST() (Thomas)</para></listitem>
8219 <listitem><para>Add pg_backend_pid() to output backend pid (Bruce)</para></listitem>
8220 <listitem><para>Add IS OF / IS NOT OF type predicate (Thomas)</para></listitem>
8221 <listitem><para>Allow bit string constants without fully-specified length (Thomas)</para></listitem>
8222 <listitem><para>Allow conversion between 8-byte integers and bit strings (Thomas)</para></listitem>
8223 <listitem><para>Implement hex literal conversion to bit string literal (Thomas)</para></listitem>
8224 <listitem><para>Allow table functions to appear in the FROM clause (Joe)</para></listitem>
8225 <listitem><para>Increase maximum number of function parameters to 32 (Bruce)</para></listitem>
8226 <listitem><para>No longer automatically create index for SERIAL column (Tom)</para></listitem>
8227 <listitem><para>Add current_database() (Rod)</para></listitem>
8228 <listitem><para>Fix cash_words() to not overflow buffer (Tom)</para></listitem>
8229 <listitem><para>Add functions replace(), split_part(), to_hex() (Joe)</para></listitem>
8230 <listitem><para>Fix LIKE for bytea as a right-hand argument (Joe)</para></listitem>
8231 <listitem><para>Prevent crashes caused by SELECT cash_out(2) (Tom)</para></listitem>
8232 <listitem><para>Fix to_char(1,'FM999.99') to return a period (Karel)</para></listitem>
8233 <listitem><para>Fix trigger/type/language functions returning OPAQUE to return proper type (Tom)</para></listitem>
8238 <title>Internationalization</title>
8240 <listitem><para>Add additional encodings: Korean (JOHAB), Thai (WIN874), Vietnamese (TCVN), Arabic (WIN1256), Simplified Chinese (GBK), Korean (UHC) (Eiji Tokuya)</para></listitem>
8241 <listitem><para>Enable locale support by default (Peter)</para></listitem>
8242 <listitem><para>Add locale variables (Peter)</para></listitem>
8243 <listitem><para>Escape byes >= 0x7f for multibyte in PQescapeBytea/PQunescapeBytea (Tatsuo)</para></listitem>
8244 <listitem><para>Add locale awareness to regular expression character classes</para></listitem>
8245 <listitem><para>Enable multibyte support by default (Tatsuo)</para></listitem>
8246 <listitem><para>Add GB18030 multibyte support (Bill Huang)</para></listitem>
8247 <listitem><para>Add CREATE/DROP CONVERSION, allowing loadable encodings (Tatsuo, Kaori)</para></listitem>
8248 <listitem><para>Add pg_conversion table (Tatsuo)</para></listitem>
8249 <listitem><para>Add SQL99 CONVERT() function (Tatsuo)</para></listitem>
8250 <listitem><para>pg_dumpall, pg_controldata, and pg_resetxlog now national-language aware (Peter)</para></listitem>
8251 <listitem><para>New and updated translations</para></listitem>
8256 <title>Server-side Languages</title>
8258 <listitem><para>Allow recursive SQL function (Peter)</para></listitem>
8259 <listitem><para>Change PL/Tcl build to use configured compiler and Makefile.shlib (Peter)</para></listitem>
8260 <listitem><para>Overhaul the PL/pgSQL FOUND variable to be more Oracle-compatible (Neil, Tom)</para></listitem>
8261 <listitem><para>Allow PL/pgSQL to handle quoted identifiers (Tom)</para></listitem>
8262 <listitem><para>Allow set-returning PL/pgSQL functions (Neil)</para></listitem>
8263 <listitem><para>Make PL/pgSQL schema-aware (Joe)</para></listitem>
8264 <listitem><para>Remove some memory leaks (Nigel J. Andrews, Tom)</para></listitem>
8271 <listitem><para>Don't lowercase psql \connect database name for 7.2.0 compatibility (Tom)</para></listitem>
8272 <listitem><para>Add psql \timing to time user queries (Greg Sabino Mullane)</para></listitem>
8273 <listitem><para>Have psql \d show index information (Greg Sabino Mullane)</para></listitem>
8274 <listitem><para>New psql \dD shows domains (Jonathan Eisler)</para></listitem>
8275 <listitem><para>Allow psql to show rules on views (Paul ?)</para></listitem>
8276 <listitem><para>Fix for psql variable substitution (Tom)</para></listitem>
8277 <listitem><para>Allow psql \d to show temporary table structure (Tom)</para></listitem>
8278 <listitem><para>Allow psql \d to show foreign keys (Rod)</para></listitem>
8279 <listitem><para>Fix \? to honor \pset pager (Bruce)</para></listitem>
8280 <listitem><para>Have psql reports its version number on startup (Tom)</para></listitem>
8281 <listitem><para>Allow \copy to specify column names (Tom)</para></listitem>
8286 <title>libpq</title>
8288 <listitem><para>Add ~/.pgpass to store host/user password combinations (Alvaro Herrera)</para></listitem>
8289 <listitem><para>Add PQunescapeBytea() function to libpq (Patrick Welche)</para></listitem>
8290 <listitem><para>Fix for sending large queries over non-blocking connections (Bernhard Herzog)</para></listitem>
8291 <listitem><para>Fix for libpq using timers on Win9X (David Ford)</para></listitem>
8292 <listitem><para>Allow libpq notify to handle servers with different-length identifiers (Tom)</para></listitem>
8293 <listitem><para>Add libpq PQescapeString() and PQescapeBytea() to Windows (Bruce)</para></listitem>
8294 <listitem><para>Fix for SSL with non-blocking connections (Jack Bates)</para></listitem>
8295 <listitem><para>Add libpq connection timeout parameter (Denis A Ustimenko)</para></listitem>
8302 <listitem><para>Allow JDBC to compile with JDK 1.4 (Dave)</para></listitem>
8303 <listitem><para>Add JDBC 3 support (Barry)</para></listitem>
8304 <listitem><para>Allows JDBC to set loglevel by adding ?loglevel=X to the connection URL (Barry)</para></listitem>
8305 <listitem><para>Add Driver.info() message that prints out the version number (Barry)</para></listitem>
8306 <listitem><para>Add updateable result sets (Raghu Nidagal, Dave)</para></listitem>
8307 <listitem><para>Add support for callable statements (Paul Bethe)</para></listitem>
8308 <listitem><para>Add query cancel capability</para></listitem>
8309 <listitem><para>Add refresh row (Dave)</para></listitem>
8310 <listitem><para>Fix MD5 encryption handling for multibyte servers (Jun Kawai)</para></listitem>
8311 <listitem><para>Add support for prepared statements (Barry)</para></listitem>
8316 <title>Miscellaneous Interfaces</title>
8318 <listitem><para>Fixed ECPG bug concerning octal numbers in single quotes (Michael)</para></listitem>
8319 <listitem><para>Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8320 <listitem><para>Improve Python interface (Elliot Lee, Andrew Johnson, Greg Copeland)</para></listitem>
8321 <listitem><para>Add libpgtcl connection close event (Gerhard Hintermayer)</para></listitem>
8322 <listitem><para>Move src/interfaces/libpq++ to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8323 <listitem><para>Move src/interfaces/odbc to http://gborg.postgresql.org (Marc)</para></listitem>
8324 <listitem><para>Move src/interfaces/libpgeasy to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8325 <listitem><para>Move src/interfaces/perl5 to http://gborg.postgresql.org (Marc, Bruce)</para></listitem>
8326 <listitem><para>Remove src/bin/pgaccess from main tree, now at http://www.pgaccess.org (Bruce)</para></listitem>
8327 <listitem><para>Add pg_on_connection_loss command to libpgtcl (Gerhard Hintermayer, Tom)</para></listitem>
8332 <title>Source Code</title>
8334 <listitem><para>Fix for parallel make (Peter)</para></listitem>
8335 <listitem><para>AIX fixes for linking Tcl (Andreas Zeugswetter)</para></listitem>
8336 <listitem><para>Allow PL/Perl to build under Cygwin (Jason Tishler)</para></listitem>
8337 <listitem><para>Improve MIPS compiles (Peter, Oliver Elphick)</para></listitem>
8338 <listitem><para>Require Autoconf version 2.53 (Peter)</para></listitem>
8339 <listitem><para>Require readline and zlib by default in configure (Peter)</para></listitem>
8340 <listitem><para>Allow Solaris to use Intimate Shared Memory (ISM), for performance (Scott Brunza, P.J. Josh Rovero)</para></listitem>
8341 <listitem><para>Always enable syslog in compile, remove --enable-syslog option (Tatsuo)</para></listitem>
8342 <listitem><para>Always enable multibyte in compile, remove --enable-multibyte option (Tatsuo)</para></listitem>
8343 <listitem><para>Always enable locale in compile, remove --enable-locale option (Peter)</para></listitem>
8344 <listitem><para>Fix for Win9x DLL creation (Magnus Naeslund)</para></listitem>
8345 <listitem><para>Fix for link() usage by WAL code on Windows, BeOS (Jason Tishler)</para></listitem>
8346 <listitem><para>Add sys/types.h to c.h, remove from main files (Peter, Bruce)</para></listitem>
8347 <listitem><para>Fix AIX hang on SMP machines (Tomoyuki Niijima)</para></listitem>
8348 <listitem><para>AIX SMP hang fix (Tomoyuki Niijima)</para></listitem>
8349 <listitem><para>Fix pre-1970 date handling on newer glibc libraries (Tom)</para></listitem>
8350 <listitem><para>Fix PowerPC SMP locking (Tom)</para></listitem>
8351 <listitem><para>Prevent gcc -ffast-math from being used (Peter, Tom)</para></listitem>
8352 <listitem><para>Bison >= 1.50 now required for developer builds</para></listitem>
8353 <listitem><para>Kerberos 5 support now builds with Heimdal (Peter)</para></listitem>
8354 <listitem><para>Add appendix in the User's Guide which lists SQL features (Thomas)</para></listitem>
8355 <listitem><para>Improve loadable module linking to use RTLD_NOW (Tom)</para></listitem>
8356 <listitem><para>New error levels WARNING, INFO, LOG, DEBUG[1-5] (Bruce)</para></listitem>
8357 <listitem><para>New src/port directory holds replaced libc functions (Peter, Bruce)</para></listitem>
8358 <listitem><para>New pg_namespace system catalog for schemas (Tom)</para></listitem>
8359 <listitem><para>Add pg_class.relnamespace for schemas (Tom)</para></listitem>
8360 <listitem><para>Add pg_type.typnamespace for schemas (Tom)</para></listitem>
8361 <listitem><para>Add pg_proc.pronamespace for schemas (Tom)</para></listitem>
8362 <listitem><para>Restructure aggregates to have pg_proc entries (Tom)</para></listitem>
8363 <listitem><para>System relations now have their own namespace, pg_* test not required (Fernando Nasser)</para></listitem>
8364 <listitem><para>Rename TOAST index names to be *_index rather than *_idx (Neil)</para></listitem>
8365 <listitem><para>Add namespaces for operators, opclasses (Tom)</para></listitem>
8366 <listitem><para>Add additional checks to server control file (Thomas)</para></listitem>
8367 <listitem><para>New Polish FAQ (Marcin Mazurek)</para></listitem>
8368 <listitem><para>Add Posix semaphore support (Tom)</para></listitem>
8369 <listitem><para>Document need for reindex (Bruce)</para></listitem>
8370 <listitem><para>Rename some internal identifiers to simplify Windows compile (Jan, Katherine Ward)</para></listitem>
8371 <listitem><para>Add documentation on computing disk space (Bruce)</para></listitem>
8372 <listitem><para>Remove KSQO from GUC (Bruce)</para></listitem>
8373 <listitem><para>Fix memory leak in rtree (Kenneth Been)</para></listitem>
8374 <listitem><para>Modify a few error messages for consistency (Bruce)</para></listitem>
8375 <listitem><para>Remove unused system table columns (Peter)</para></listitem>
8376 <listitem><para>Make system columns NOT NULL where appropriate (Tom)</para></listitem>
8377 <listitem><para>Clean up use of sprintf in favor of snprintf() (Neil, Jukka Holappa)</para></listitem>
8378 <listitem><para>Remove OPAQUE and create specific subtypes (Tom)</para></listitem>
8379 <listitem><para>Cleanups in array internal handling (Joe, Tom)</para></listitem>
8380 <listitem><para>Disallow pg_atoi('') (Bruce)</para></listitem>
8381 <listitem><para>Remove parameter wal_files because WAL files are now recycled (Bruce)</para></listitem>
8382 <listitem><para>Add version numbers to heap pages (Tom)</para></listitem>
8387 <title>Contrib</title>
8389 <listitem><para>Allow inet arrays in /contrib/array (Neil)</para></listitem>
8390 <listitem><para>GiST fixes (Teodor Sigaev, Neil)</para></listitem>
8391 <listitem><para>Upgrade /contrib/mysql</para></listitem>
8392 <listitem><para>Add /contrib/dbsize which shows table sizes without vacuum (Peter)</para></listitem>
8393 <listitem><para>Add /contrib/intagg, integer aggregator routines (mlw)</para></listitem>
8394 <listitem><para>Improve /contrib/oid2name (Neil, Bruce)</para></listitem>
8395 <listitem><para>Improve /contrib/tsearch (Oleg, Teodor Sigaev)</para></listitem>
8396 <listitem><para>Cleanups of /contrib/rserver (Alexey V. Borzov)</para></listitem>
8397 <listitem><para>Update /contrib/oracle conversion utility (Gilles Darold)</para></listitem>
8398 <listitem><para>Update /contrib/dblink (Joe)</para></listitem>
8399 <listitem><para>Improve options supported by /contrib/vacuumlo (Mario Weilguni)</para></listitem>
8400 <listitem><para>Improvements to /contrib/intarray (Oleg, Teodor Sigaev, Andrey Oktyabrski)</para></listitem>
8401 <listitem><para>Add /contrib/reindexdb utility (Shaun Thomas)</para></listitem>
8402 <listitem><para>Add indexing to /contrib/isbn_issn (Dan Weston)</para></listitem>
8403 <listitem><para>Add /contrib/dbmirror (Steven Singer)</para></listitem>
8404 <listitem><para>Improve /contrib/pgbench (Neil)</para></listitem>
8405 <listitem><para>Add /contrib/tablefunc table function examples (Joe)</para></listitem>
8406 <listitem><para>Add /contrib/ltree data type for tree structures (Teodor Sigaev, Oleg Bartunov)</para></listitem>
8407 <listitem><para>Move /contrib/pg_controldata, pg_resetxlog into main tree (Bruce)</para></listitem>
8408 <listitem><para>Fixes to /contrib/cube (Bruno Wolff)</para></listitem>
8409 <listitem><para>Improve /contrib/fulltextindex (Christopher)</para></listitem>
8416 <sect1 id="release-7-2-8">
8417 <title>Release 7.2.8</title>
8420 <title>Release date</title>
8421 <simpara>2005-05-09</simpara>
8425 This release contains a variety of fixes from 7.2.7, including one
8426 security-related issue.
8430 <title>Migration to version 7.2.8</title>
8433 A dump/restore is not required for those running 7.2.X.
8438 <title>Changes</title>
8441 <listitem><para>Repair ancient race condition that allowed a transaction to be
8442 seen as committed for some purposes (eg SELECT FOR UPDATE) slightly sooner
8443 than for other purposes</para>
8444 <para>This is an extremely serious bug since it could lead to apparent
8445 data inconsistencies being briefly visible to applications.</para></listitem>
8446 <listitem><para>Repair race condition between relation extension and
8448 <para>This could theoretically have caused loss of a page's worth of
8449 freshly-inserted data, although the scenario seems of very low probability.
8450 There are no known cases of it having caused more than an Assert failure.
8452 <listitem><para>Fix <function>EXTRACT(EPOCH)</> for
8453 <type>TIME WITH TIME ZONE</> values</para></listitem>
8454 <listitem><para>Additional buffer overrun checks in plpgsql
8455 (Neil)</para></listitem>
8456 <listitem><para>Fix pg_dump to dump index names and trigger names containing
8457 <literal>%</> correctly (Neil)</para></listitem>
8458 <listitem><para>Prevent <function>to_char(interval)</> from dumping core for
8459 month-related formats</para></listitem>
8460 <listitem><para>Fix <filename>contrib/pgcrypto</> for newer OpenSSL builds
8461 (Marko Kreen)</para></listitem>
8467 <sect1 id="release-7-2-7">
8468 <title>Release 7.2.7</title>
8471 <title>Release date</title>
8472 <simpara>2005-01-31</simpara>
8476 This release contains a variety of fixes from 7.2.6, including several
8477 security-related issues.
8481 <title>Migration to version 7.2.7</title>
8484 A dump/restore is not required for those running 7.2.X.
8489 <title>Changes</title>
8492 <listitem><para>Disallow <command>LOAD</> to non-superusers</para>
8494 On platforms that will automatically execute initialization functions of a
8495 shared library (this includes at least Windows and ELF-based Unixen),
8496 <command>LOAD</> can be used to make the server execute arbitrary code.
8497 Thanks to NGS Software for reporting this.</para></listitem>
8498 <listitem><para>Add needed STRICT marking to some contrib functions (Kris
8499 Jurka)</para></listitem>
8500 <listitem><para>Avoid buffer overrun when plpgsql cursor declaration has too
8501 many parameters (Neil)</para></listitem>
8502 <listitem><para>Fix planning error for FULL and RIGHT outer joins</para>
8504 The result of the join was mistakenly supposed to be sorted the same as the
8505 left input. This could not only deliver mis-sorted output to the user, but
8506 in case of nested merge joins could give outright wrong answers.
8508 <listitem><para>Fix display of negative intervals in SQL and GERMAN
8509 datestyles</para></listitem>
8515 <sect1 id="release-7-2-6">
8516 <title>Release 7.2.6</title>
8519 <title>Release date</title>
8520 <simpara>2004-10-22</simpara>
8524 This release contains a variety of fixes from 7.2.5.
8529 <title>Migration to version 7.2.6</title>
8532 A dump/restore is not required for those running 7.2.X.
8537 <title>Changes</title>
8540 <listitem><para>Repair possible failure to update hint bits on disk</para>
8542 Under rare circumstances this oversight could lead to
8543 <quote>could not access transaction status</> failures, which qualifies
8544 it as a potential-data-loss bug.
8546 <listitem><para>Ensure that hashed outer join does not miss tuples</para>
8548 Very large left joins using a hash join plan could fail to output unmatched
8549 left-side rows given just the right data distribution.
8551 <listitem><para>Disallow running pg_ctl as root</para>
8553 This is to guard against any possible security issues.
8555 <listitem><para>Avoid using temp files in /tmp in make_oidjoins_check</para>
8557 This has been reported as a security issue, though it's hardly worthy of
8558 concern since there is no reason for non-developers to use this script anyway.
8560 <listitem><para>Update to newer versions of Bison</para></listitem>
8566 <sect1 id="release-7-2-5">
8567 <title>Release 7.2.5</title>
8570 <title>Release date</title>
8571 <simpara>2004-08-16</simpara>
8575 This release contains a variety of fixes from 7.2.4.
8580 <title>Migration to version 7.2.5</title>
8583 A dump/restore is not required for those running 7.2.X.
8588 <title>Changes</title>
8591 <listitem><para>Prevent possible loss of committed transactions during crash</para>
8593 Due to insufficient interlocking between transaction commit and checkpointing,
8594 it was possible for transactions committed just before the most recent
8595 checkpoint to be lost, in whole or in part, following a database crash and
8596 restart. This is a serious bug that has existed
8597 since <productname>PostgreSQL</productname> 7.1.
8599 <listitem><para>Fix corner case for btree search in parallel with first root page split</para></listitem>
8600 <listitem><para>Fix buffer overrun in <function>to_ascii</function> (Guido Notari)</para></listitem>
8601 <listitem><para>Fix core dump in deadlock detection on machines where char is unsigned</para></listitem>
8602 <listitem><para>Fix failure to respond to <command>pg_ctl stop -m fast</command> after Async_NotifyHandler runs</para></listitem>
8603 <listitem><para>Repair memory leaks in pg_dump</para></listitem>
8604 <listitem><para>Avoid conflict with system definition of <function>isblank()</function> function or macro</para></listitem>
8609 <sect1 id="release-7-2-4">
8610 <title>Release 7.2.4</title>
8613 <title>Release date</title>
8614 <simpara>2003-01-30</simpara>
8618 This release contains a variety of fixes for version 7.2.3,
8619 including fixes to prevent possible data loss.
8623 <title>Migration to version 7.2.4</title>
8626 A dump/restore is <emphasis>not</emphasis> required for those
8627 running version 7.2.*.
8632 <title>Changes</title>
8635 <listitem><para>Fix some additional cases of VACUUM "No one parent tuple was found" error</para></listitem>
8636 <listitem><para>Prevent VACUUM from being called inside a function (Bruce)</para></listitem>
8637 <listitem><para>Ensure pg_clog updates are sync'd to disk before marking checkpoint complete</para></listitem>
8638 <listitem><para>Avoid integer overflow during large hash joins</para></listitem>
8639 <listitem><para>Make GROUP commands work when pg_group.grolist is large enough to be toasted</para></listitem>
8640 <listitem><para>Fix errors in datetime tables; some timezone names weren't being recognized</para></listitem>
8641 <listitem><para>Fix integer overflows in circle_poly(), path_encode(), path_add() (Neil)</para></listitem>
8642 <listitem><para>Repair long-standing logic errors in lseg_eq(), lseg_ne(), lseg_center()</para></listitem>
8648 <sect1 id="release-7-2-3">
8649 <title>Release 7.2.3</title>
8652 <title>Release date</title>
8653 <simpara>2002-10-01</simpara>
8657 This release contains a variety of fixes for version 7.2.2,
8658 including fixes to prevent possible data loss.
8662 <title>Migration to version 7.2.3</title>
8665 A dump/restore is <emphasis>not</emphasis> required for those
8666 running version 7.2.*.
8671 <title>Changes</title>
8674 <listitem><para>Prevent possible compressed transaction log loss (Tom)</para></listitem>
8675 <listitem><para>Prevent non-superuser from increasing most recent vacuum info (Tom)</para></listitem>
8676 <listitem><para>Handle pre-1970 date values in newer versions of glibc (Tom)</para></listitem>
8677 <listitem><para>Fix possible hang during server shutdown</para></listitem>
8678 <listitem><para>Prevent spinlock hangs on SMP PPC machines (Tomoyuki Niijima)</para></listitem>
8679 <listitem><para>Fix <application>pg_dump</> to properly dump FULL JOIN USING (Tom)</para></listitem>
8685 <sect1 id="release-7-2-2">
8686 <title>Release 7.2.2</title>
8689 <title>Release date</title>
8690 <simpara>2002-08-23</simpara>
8694 This release contains a variety of fixes for version 7.2.1.
8698 <title>Migration to version 7.2.2</title>
8701 A dump/restore is <emphasis>not</emphasis> required for those
8702 running version 7.2.*.
8707 <title>Changes</title>
8710 <listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
8711 <listitem><para>Fix for compressed transaction log id wraparound (Tom)</para></listitem>
8712 <listitem><para>Fix PQescapeBytea/PQunescapeBytea so that they handle bytes > 0x7f (Tatsuo)</para></listitem>
8713 <listitem><para>Fix for psql and <application>pg_dump</> crashing when invoked with non-existent long options (Tatsuo)</para></listitem>
8714 <listitem><para>Fix crash when invoking geometric operators (Tom)</para></listitem>
8715 <listitem><para>Allow OPEN cursor(args) (Tom)</para></listitem>
8716 <listitem><para>Fix for rtree_gist index build (Teodor)</para></listitem>
8717 <listitem><para>Fix for dumping user-defined aggregates (Tom)</para></listitem>
8718 <listitem><para>contrib/intarray fixes (Oleg)</para></listitem>
8719 <listitem><para>Fix for complex UNION/EXCEPT/INTERSECT queries using parens (Tom)</para></listitem>
8720 <listitem><para>Fix to pg_convert (Tatsuo)</para></listitem>
8721 <listitem><para>Fix for crash with long DATA strings (Thomas, Neil)</para></listitem>
8722 <listitem><para>Fix for repeat(), lpad(), rpad() and long strings (Neil)</para></listitem>
8728 <sect1 id="release-7-2-1">
8729 <title>Release 7.2.1</title>
8732 <title>Release date</title>
8733 <simpara>2002-03-21</simpara>
8737 This release contains a variety of fixes for version 7.2.
8741 <title>Migration to version 7.2.1</title>
8744 A dump/restore is <emphasis>not</emphasis> required for those
8745 running version 7.2.
8750 <title>Changes</title>
8753 <listitem><para>Ensure that sequence counters do not go backwards after a crash (Tom)</para></listitem>
8754 <listitem><para>Fix pgaccess kanji-conversion key binding (Tatsuo)</para></listitem>
8755 <listitem><para>Optimizer improvements (Tom)</para></listitem>
8756 <listitem><para>Cash I/O improvements (Tom)</para></listitem>
8757 <listitem><para>New Russian FAQ</para></listitem>
8758 <listitem><para>Compile fix for missing AuthBlockSig (Heiko)</para></listitem>
8759 <listitem><para>Additional time zones and time zone fixes (Thomas)</para></listitem>
8760 <listitem><para>Allow psql \connect to handle mixed case database and user names (Tom)</para></listitem>
8761 <listitem><para>Return proper OID on command completion even with ON INSERT rules (Tom)</para></listitem>
8762 <listitem><para>Allow COPY FROM to use 8-bit DELIMITERS (Tatsuo)</para></listitem>
8763 <listitem><para>Fix bug in extract/date_part for milliseconds/microseconds (Tatsuo)</para></listitem>
8764 <listitem><para>Improve handling of multiple UNIONs with different lengths (Tom)</para></listitem>
8765 <listitem><para>contrib/btree_gist improvements (Teodor Sigaev)</para></listitem>
8766 <listitem><para>contrib/tsearch dictionary improvements, see README.tsearch for an additional installation step (Thomas T. Thai, Teodor Sigaev)</para></listitem>
8767 <listitem><para>Fix for array subscripts handling (Tom)</para></listitem>
8768 <listitem><para>Allow EXECUTE of "CREATE TABLE AS ... SELECT" in PL/pgSQL (Tom)</para></listitem>
8774 <sect1 id="release-7-2">
8775 <title>Release 7.2</title>
8778 <title>Release date</title>
8779 <simpara>2002-02-04</simpara>
8783 <title>Overview</title>
8786 This release improves <productname>PostgreSQL</> for use in
8787 high-volume applications.
8791 Major changes in this release:
8799 Vacuuming no longer locks tables, thus allowing normal user
8800 access during the vacuum. A new <command>VACUUM FULL</>
8801 command does old-style vacuum by locking the table and
8802 shrinking the on-disk copy of the table.
8808 <term>Transactions</term>
8811 There is no longer a problem with installations that exceed
8812 four billion transactions.
8821 OIDs are now optional. Users can now create tables without
8822 OIDs for cases where OID usage is excessive.
8828 <term>Optimizer</term>
8831 The system now computes histogram column statistics during
8832 <command>ANALYZE</>, allowing much better optimizer choices.
8838 <term>Security</term>
8841 A new MD5 encryption option allows more secure storage and
8842 transfer of passwords. A new Unix-domain socket
8843 authentication option is available on Linux and BSD systems.
8849 <term>Statistics</term>
8852 Administrators can use the new table access statistics module
8853 to get fine-grained information about table and index usage.
8859 <term>Internationalization</term>
8862 Program and library messages can now be displayed in several
8872 <title>Migration to version 7.2</title>
8875 A dump/restore using <command>pg_dump</command> is required for
8876 those wishing to migrate data from any previous release.
8880 Observe the following incompatibilities:
8886 The semantics of the <command>VACUUM</command> command have
8887 changed in this release. You may wish to update your
8888 maintenance procedures accordingly.
8894 In this release, comparisons using <literal>= NULL</literal>
8895 will always return false (or NULL, more precisely). Previous
8896 releases automatically transformed this syntax to <literal>IS
8897 NULL</literal>. The old behavior can be re-enabled using a
8898 <filename>postgresql.conf</filename> parameter.
8904 The <filename>pg_hba.conf</> and <filename>pg_ident.conf</>
8905 configuration is now only reloaded after receiving a
8906 <systemitem>SIGHUP</> signal, not with each connection.
8912 The function <filename>octet_length()</> now returns the uncompressed data length.
8918 The date/time value <literal>'current'</literal> is no longer
8919 available. You will need to rewrite your applications.
8925 The <literal>timestamp()</literal>, <literal>time()</literal>,
8926 and <literal>interval()</literal> functions are no longer
8927 available. Instead of <literal>timestamp()</literal>, use
8928 <literal>timestamp 'string'</literal> or <literal>CAST</literal>.
8935 The <literal>SELECT ... LIMIT #,#</literal> syntax will be removed
8936 in the next release. You should change your queries to use
8937 separate LIMIT and OFFSET clauses, e.g. <literal>LIMIT 10 OFFSET
8943 <title>Changes</title>
8946 <title>Server Operation</title>
8948 <listitem><para>Create temporary files in a separate directory (Bruce)</para></listitem>
8949 <listitem><para>Delete orphaned temporary files on postmaster startup (Bruce)</para></listitem>
8950 <listitem><para>Added unique indexes to some system tables (Tom)</para></listitem>
8951 <listitem><para>System table operator reorganization (Oleg Bartunov, Teodor Sigaev, Tom)</para></listitem>
8952 <listitem><para>Renamed pg_log to pg_clog (Tom)</para></listitem>
8953 <listitem><para>Enable SIGTERM, SIGQUIT to kill backends (Jan)</para></listitem>
8954 <listitem><para>Removed compile-time limit on number of backends (Tom)</para></listitem>
8955 <listitem><para>Better cleanup for semaphore resource failure (Tatsuo, Tom)</para></listitem>
8956 <listitem><para>Allow safe transaction ID wraparound (Tom)</para></listitem>
8957 <listitem><para>Removed OIDs from some system tables (Tom)</para></listitem>
8958 <listitem><para>Removed "triggered data change violation" error check (Tom)</para></listitem>
8959 <listitem><para>SPI portal creation of prepared/saved plans (Jan)</para></listitem>
8960 <listitem><para>Allow SPI column functions to work for system columns (Tom)</para></listitem>
8961 <listitem><para>Long value compression improvement (Tom)</para></listitem>
8962 <listitem><para>Statistics collector for table, index access (Jan)</para></listitem>
8963 <listitem><para>Truncate extra-long sequence names to a reasonable value (Tom)</para></listitem>
8964 <listitem><para>Measure transaction times in milliseconds (Thomas)</para></listitem>
8965 <listitem><para>Fix TID sequential scans (Hiroshi)</para></listitem>
8966 <listitem><para>Superuser ID now fixed at 1 (Peter E)</para></listitem>
8967 <listitem><para>New pg_ctl "reload" option (Tom)</para></listitem>
8972 <title>Performance</title>
8974 <listitem><para>Optimizer improvements (Tom)</para></listitem>
8975 <listitem><para>New histogram column statistics for optimizer (Tom)</para></listitem>
8976 <listitem><para>Reuse write-ahead log files rather than discarding them (Tom)</para></listitem>
8977 <listitem><para>Cache improvements (Tom)</para></listitem>
8978 <listitem><para>IS NULL, IS NOT NULL optimizer improvement (Tom)</para></listitem>
8979 <listitem><para>Improve lock manager to reduce lock contention (Tom)</para></listitem>
8980 <listitem><para>Keep relcache entries for index access support functions (Tom)</para></listitem>
8981 <listitem><para>Allow better selectivity with NaN and infinities in NUMERIC (Tom)</para></listitem>
8982 <listitem><para>R-tree performance improvements (Kenneth Been)</para></listitem>
8983 <listitem><para>B-tree splits more efficient (Tom)</para></listitem>
8988 <title>Privileges</title>
8990 <listitem><para>Change UPDATE, DELETE privileges to be distinct (Peter E)</para></listitem>
8991 <listitem><para>New REFERENCES, TRIGGER privileges (Peter E)</para></listitem>
8992 <listitem><para>Allow GRANT/REVOKE to/from more than one user at a time (Peter E)</para></listitem>
8993 <listitem><para>New has_table_privilege() function (Joe Conway)</para></listitem>
8994 <listitem><para>Allow non-superuser to vacuum database (Tom)</para></listitem>
8995 <listitem><para>New SET SESSION AUTHORIZATION command (Peter E)</para></listitem>
8996 <listitem><para>Fix bug in privilege modifications on newly created tables (Tom)</para></listitem>
8997 <listitem><para>Disallow access to pg_statistic for non-superuser, add user-accessible views (Tom)</para></listitem>
9002 <title>Client Authentication</title>
9004 <listitem><para>Fork postmaster before doing authentication to prevent hangs (Peter E)</para></listitem>
9005 <listitem><para>Add ident authentication over Unix domain sockets on Linux, *BSD (Helge Bahmann, Oliver Elphick, Teodor Sigaev, Bruce)</para></listitem>
9006 <listitem><para>Add a password authentication method that uses MD5 encryption (Bruce)</para></listitem>
9007 <listitem><para>Allow encryption of stored passwords using MD5 (Bruce)</para></listitem>
9008 <listitem><para>PAM authentication (Dominic J. Eidson)</para></listitem>
9009 <listitem><para>Load pg_hba.conf and pg_ident.conf only on startup and SIGHUP (Bruce)</para></listitem>
9014 <title>Server Configuration</title>
9016 <listitem><para>Interpretation of some time zone abbreviations as Australian rather than North American now settable at run time (Bruce)</para></listitem>
9017 <listitem><para>New parameter to set default transaction isolation level (Peter E)</para></listitem>
9018 <listitem><para>New parameter to enable conversion of "expr = NULL" into "expr IS NULL", off by default (Peter E)</para></listitem>
9019 <listitem><para>New parameter to control memory usage by VACUUM (Tom)</para></listitem>
9020 <listitem><para>New parameter to set client authentication timeout (Tom)</para></listitem>
9021 <listitem><para>New parameter to set maximum number of open files (Tom)</para></listitem>
9026 <title>Queries</title>
9028 <listitem><para>Statements added by INSERT rules now execute after the INSERT (Jan)</para></listitem>
9029 <listitem><para>Prevent unadorned relation names in target list (Bruce)</para></listitem>
9030 <listitem><para>NULLs now sort after all normal values in ORDER BY (Tom)</para></listitem>
9031 <listitem><para>New IS UNKNOWN, IS NOT UNKNOWN Boolean tests (Tom)</para></listitem>
9032 <listitem><para>New SHARE UPDATE EXCLUSIVE lock mode (Tom)</para></listitem>
9033 <listitem><para>New EXPLAIN ANALYZE command that shows run times and row counts (Martijn van Oosterhout)</para></listitem>
9034 <listitem><para>Fix problem with LIMIT and subqueries (Tom)</para></listitem>
9035 <listitem><para>Fix for LIMIT, DISTINCT ON pushed into subqueries (Tom)</para></listitem>
9036 <listitem><para>Fix nested EXCEPT/INTERSECT (Tom)</para></listitem>
9041 <title>Schema Manipulation</title>
9043 <listitem><para>Fix SERIAL in temporary tables (Bruce)</para></listitem>
9044 <listitem><para>Allow temporary sequences (Bruce)</para></listitem>
9045 <listitem><para>Sequences now use int8 internally (Tom)</para></listitem>
9046 <listitem><para>New SERIAL8 creates int8 columns with sequences, default still SERIAL4 (Tom)</para></listitem>
9047 <listitem><para>Make OIDs optional using WITHOUT OIDS (Tom)</para></listitem>
9048 <listitem><para>Add %TYPE syntax to CREATE TYPE (Ian Lance Taylor)</para></listitem>
9049 <listitem><para>Add ALTER TABLE / DROP CONSTRAINT for CHECK constraints (Christopher Kings-Lynne)</para></listitem>
9050 <listitem><para>New CREATE OR REPLACE FUNCTION to alter existing function (preserving the function OID) (Gavin Sherry)</para></listitem>
9051 <listitem><para>Add ALTER TABLE / ADD [ UNIQUE | PRIMARY ] (Christopher Kings-Lynne)</para></listitem>
9052 <listitem><para>Allow column renaming in views</para></listitem>
9053 <listitem><para>Make ALTER TABLE / RENAME COLUMN update column names of indexes (Brent Verner)</para></listitem>
9054 <listitem><para>Fix for ALTER TABLE / ADD CONSTRAINT ... CHECK with inherited tables (Stephan Szabo)</para></listitem>
9055 <listitem><para>ALTER TABLE RENAME update foreign-key trigger arguments correctly (Brent Verner)</para></listitem>
9056 <listitem><para>DROP AGGREGATE and COMMENT ON AGGREGATE now accept an aggtype (Tom)</para></listitem>
9057 <listitem><para>Add automatic return type data casting for SQL functions (Tom)</para></listitem>
9058 <listitem><para>Allow GiST indexes to handle NULLs and multikey indexes (Oleg Bartunov, Teodor Sigaev, Tom)</para></listitem>
9059 <listitem><para>Enable partial indexes (Martijn van Oosterhout)</para></listitem>
9064 <title>Utility Commands</title>
9066 <listitem><para>Add RESET ALL, SHOW ALL (Marko Kreen)</para></listitem>
9067 <listitem><para>CREATE/ALTER USER/GROUP now allow options in any order (Vince)</para></listitem>
9068 <listitem><para>Add LOCK A, B, C functionality (Neil Padgett)</para></listitem>
9069 <listitem><para>New ENCRYPTED/UNENCRYPTED option to CREATE/ALTER USER (Bruce)</para></listitem>
9070 <listitem><para>New light-weight VACUUM does not lock table; old semantics are available as VACUUM FULL (Tom)</para></listitem>
9071 <listitem><para>Disable COPY TO/FROM on views (Bruce)</para></listitem>
9072 <listitem><para>COPY DELIMITERS string must be exactly one character (Tom)</para></listitem>
9073 <listitem><para>VACUUM warning about index tuples fewer than heap now only appears when appropriate (Martijn van Oosterhout)</para></listitem>
9074 <listitem><para>Fix privilege checks for CREATE INDEX (Tom)</para></listitem>
9075 <listitem><para>Disallow inappropriate use of CREATE/DROP INDEX/TRIGGER/VIEW (Tom)</para></listitem>
9080 <title>Data Types and Functions</title>
9082 <listitem><para>SUM(), AVG(), COUNT() now uses int8 internally for speed (Tom)</para></listitem>
9083 <listitem><para>Add convert(), convert2() (Tatsuo)</para></listitem>
9084 <listitem><para>New function bit_length() (Peter E)</para></listitem>
9085 <listitem><para>Make the "n" in CHAR(n)/VARCHAR(n) represents letters, not bytes (Tatsuo)</para></listitem>
9086 <listitem><para>CHAR(), VARCHAR() now reject strings that are too long (Peter E)</para></listitem>
9087 <listitem><para>BIT VARYING now rejects bit strings that are too long (Peter E)</para></listitem>
9088 <listitem><para>BIT now rejects bit strings that do not match declared size (Peter E)</para></listitem>
9089 <listitem><para>INET, CIDR text conversion functions (Alex Pilosov)</para></listitem>
9090 <listitem><para>INET, CIDR operators << and <<= indexable (Alex Pilosov)</para></listitem>
9091 <listitem><para>Bytea \### now requires valid three digit octal number</para></listitem>
9092 <listitem><para>Bytea comparison improvements, now supports =, <>, >, >=, <, and <=</para></listitem>
9093 <listitem><para>Bytea now supports B-tree indexes</para></listitem>
9094 <listitem><para>Bytea now supports LIKE, LIKE...ESCAPE, NOT LIKE, NOT LIKE...ESCAPE</para></listitem>
9095 <listitem><para>Bytea now supports concatenation</para></listitem>
9096 <listitem><para>New bytea functions: position, substring, trim, btrim, and length</para></listitem>
9097 <listitem><para>New encode() function mode, "escaped", converts minimally escaped bytea to/from text</para></listitem>
9098 <listitem><para>Add pg_database_encoding_max_length() (Tatsuo)</para></listitem>
9099 <listitem><para>Add pg_client_encoding() function (Tatsuo)</para></listitem>
9100 <listitem><para>now() returns time with millisecond precision (Thomas)</para></listitem>
9101 <listitem><para>New TIMESTAMP WITHOUT TIMEZONE data type (Thomas)</para></listitem>
9102 <listitem><para>Add ISO date/time specification with "T", yyyy-mm-ddThh:mm:ss (Thomas)</para></listitem>
9103 <listitem><para>New xid/int comparison functions (Hiroshi)</para></listitem>
9104 <listitem><para>Add precision to TIME, TIMESTAMP, and INTERVAL data types (Thomas)</para></listitem>
9105 <listitem><para>Modify type coercion logic to attempt binary-compatible functions first (Tom)</para></listitem>
9106 <listitem><para>New encode() function installed by default (Marko Kreen)</para></listitem>
9107 <listitem><para>Improved to_*() conversion functions (Karel Zak)</para></listitem>
9108 <listitem><para>Optimize LIKE/ILIKE when using single-byte encodings (Tatsuo)</para></listitem>
9109 <listitem><para>New functions in contrib/pgcrypto: crypt(), hmac(), encrypt(), gen_salt() (Marko Kreen)</para></listitem>
9110 <listitem><para>Correct description of translate() function (Bruce)</para></listitem>
9111 <listitem><para>Add INTERVAL argument for SET TIME ZONE (Thomas)</para></listitem>
9112 <listitem><para>Add INTERVAL YEAR TO MONTH (etc.) syntax (Thomas)</para></listitem>
9113 <listitem><para>Optimize length functions when using single-byte encodings (Tatsuo)</para></listitem>
9114 <listitem><para>Fix path_inter, path_distance, path_length, dist_ppath to handle closed paths (Curtis Barrett, Tom)</para></listitem>
9115 <listitem><para>octet_length(text) now returns non-compressed length (Tatsuo, Bruce)</para></listitem>
9116 <listitem><para>Handle "July" full name in date/time literals (Greg Sabino Mullane)</para></listitem>
9117 <listitem><para>Some datatype() function calls now evaluated differently</para></listitem>
9118 <listitem><para>Add support for Julian and ISO time specifications (Thomas)</para></listitem>
9123 <title>Internationalization</title>
9125 <listitem><para>National language support in psql, <application>pg_dump</>, libpq, and server (Peter E)</para></listitem>
9126 <listitem><para>Message translations in Chinese (simplified, traditional), Czech, French, German, Hungarian, Russian, Swedish (Peter E, Serguei A. Mokhov, Karel Zak, Weiping He, Zhenbang Wei, Kovacs Zoltan)</para></listitem>
9127 <listitem><para>Make trim, ltrim, rtrim, btrim, lpad, rpad, translate multibyte aware (Tatsuo)</para></listitem>
9128 <listitem><para>Add LATIN5,6,7,8,9,10 support (Tatsuo)</para></listitem>
9129 <listitem><para>Add ISO 8859-5,6,7,8 support (Tatsuo)</para></listitem>
9130 <listitem><para>Correct LATIN5 to mean ISO-8859-9, not ISO-8859-5 (Tatsuo)</para></listitem>
9131 <listitem><para>Make mic2ascii() non-ASCII aware (Tatsuo)</para></listitem>
9132 <listitem><para>Reject invalid multibyte character sequences (Tatsuo)</para></listitem>
9137 <title><application>PL/pgSQL</></title>
9139 <listitem><para>Now uses portals for SELECT loops, allowing huge result sets (Jan)</para></listitem>
9140 <listitem><para>CURSOR and REFCURSOR support (Jan)</para></listitem>
9141 <listitem><para>Can now return open cursors (Jan)</para></listitem>
9142 <listitem><para>Add ELSEIF (Klaus Reger)</para></listitem>
9143 <listitem><para>Improve PL/pgSQL error reporting, including location of error (Tom)</para></listitem>
9144 <listitem><para>Allow IS or FOR key words in cursor declaration, for compatibility (Bruce)</para></listitem>
9145 <listitem><para>Fix for SELECT ... FOR UPDATE (Tom)</para></listitem>
9146 <listitem><para>Fix for PERFORM returning multiple rows (Tom)</para></listitem>
9147 <listitem><para>Make PL/pgSQL use the server's type coercion code (Tom)</para></listitem>
9148 <listitem><para>Memory leak fix (Jan, Tom)</para></listitem>
9149 <listitem><para>Make trailing semicolon optional (Tom)</para></listitem>
9154 <title>PL/Perl</title>
9156 <listitem><para>New untrusted PL/Perl (Alex Pilosov)</para></listitem>
9157 <listitem><para>PL/Perl is now built on some platforms even if libperl is not shared (Peter E)</para></listitem>
9162 <title>PL/Tcl</title>
9164 <listitem><para>Now reports errorInfo (Vsevolod Lobko)</para></listitem>
9165 <listitem><para>Add spi_lastoid function (bob@redivi.com)</para></listitem>
9170 <title>PL/Python</title>
9172 <listitem><para>...is new (Andrew Bosma)</para></listitem>
9177 <title><application>psql</></title>
9179 <listitem><para>\d displays indexes in unique, primary groupings (Christopher Kings-Lynne)</para></listitem>
9180 <listitem><para>Allow trailing semicolons in backslash commands (Greg Sabino Mullane)</para></listitem>
9181 <listitem><para>Read password from /dev/tty if possible</para></listitem>
9182 <listitem><para>Force new password prompt when changing user and database (Tatsuo, Tom)</para></listitem>
9183 <listitem><para>Format the correct number of columns for Unicode (Patrice)</para></listitem>
9188 <title><application>libpq</></title>
9190 <listitem><para>New function PQescapeString() to escape quotes in command strings (Florian Weimer)</para></listitem>
9191 <listitem><para>New function PQescapeBytea() escapes binary strings for use as SQL string literals</para></listitem>
9198 <listitem><para>Return OID of INSERT (Ken K)</para></listitem>
9199 <listitem><para>Handle more data types (Ken K)</para></listitem>
9200 <listitem><para>Handle single quotes and newlines in strings (Ken K)</para></listitem>
9201 <listitem><para>Handle NULL variables (Ken K)</para></listitem>
9202 <listitem><para>Fix for time zone handling (Barry Lind)</para></listitem>
9203 <listitem><para>Improved Druid support</para></listitem>
9204 <listitem><para>Allow eight-bit characters with non-multibyte server (Barry Lind)</para></listitem>
9205 <listitem><para>Support BIT, BINARY types (Ned Wolpert)</para></listitem>
9206 <listitem><para>Reduce memory usage (Michael Stephens, Dave Cramer)</para></listitem>
9207 <listitem><para>Update DatabaseMetaData (Peter E)</para></listitem>
9208 <listitem><para>Add DatabaseMetaData.getCatalogs() (Peter E)</para></listitem>
9209 <listitem><para>Encoding fixes (Anders Bengtsson)</para></listitem>
9210 <listitem><para>Get/setCatalog methods (Jason Davies)</para></listitem>
9211 <listitem><para>DatabaseMetaData.getColumns() now returns column defaults (Jason Davies)</para></listitem>
9212 <listitem><para>DatabaseMetaData.getColumns() performance improvement (Jeroen van Vianen)</para></listitem>
9213 <listitem><para>Some JDBC1 and JDBC2 merging (Anders Bengtsson)</para></listitem>
9214 <listitem><para>Transaction performance improvements (Barry Lind)</para></listitem>
9215 <listitem><para>Array fixes (Greg Zoller)</para></listitem>
9216 <listitem><para>Serialize addition </para></listitem>
9217 <listitem><para>Fix batch processing (Rene Pijlman)</para></listitem>
9218 <listitem><para>ExecSQL method reorganization (Anders Bengtsson)</para></listitem>
9219 <listitem><para>GetColumn() fixes (Jeroen van Vianen)</para></listitem>
9220 <listitem><para>Fix isWriteable() function (Rene Pijlman)</para></listitem>
9221 <listitem><para>Improved passage of JDBC2 conformance tests (Rene Pijlman)</para></listitem>
9222 <listitem><para>Add bytea type capability (Barry Lind)</para></listitem>
9223 <listitem><para>Add isNullable() (Rene Pijlman)</para></listitem>
9224 <listitem><para>JDBC date/time test suite fixes (Liam Stewart)</para></listitem>
9225 <listitem><para>Fix for SELECT 'id' AS xxx FROM table (Dave Cramer)</para></listitem>
9226 <listitem><para>Fix DatabaseMetaData to show precision properly (Mark Lillywhite)</para></listitem>
9227 <listitem><para>New getImported/getExported keys (Jason Davies)</para></listitem>
9228 <listitem><para>MD5 password encryption support (Jeremy Wohl)</para></listitem>
9229 <listitem><para>Fix to actually use type cache (Ned Wolpert)</para></listitem>
9236 <listitem><para>Remove query size limit (Hiroshi)</para></listitem>
9237 <listitem><para>Remove text field size limit (Hiroshi)</para></listitem>
9238 <listitem><para>Fix for SQLPrimaryKeys in multibyte mode (Hiroshi)</para></listitem>
9239 <listitem><para>Allow ODBC procedure calls (Hiroshi)</para></listitem>
9240 <listitem><para>Improve boolean handing (Aidan Mountford)</para></listitem>
9241 <listitem><para>Most configuration options now settable via DSN (Hiroshi)</para></listitem>
9242 <listitem><para>Multibyte, performance fixes (Hiroshi)</para></listitem>
9243 <listitem><para>Allow driver to be used with iODBC or unixODBC (Peter E)</para></listitem>
9244 <listitem><para>MD5 password encryption support (Bruce)</para></listitem>
9245 <listitem><para>Add more compatibility functions to odbc.sql (Peter E)</para></listitem>
9250 <title><application>ECPG</></title>
9252 <listitem><para>EXECUTE ... INTO implemented (Christof Petig)</para></listitem>
9253 <listitem><para>Multiple row descriptor support (e.g. CARDINALITY) (Christof Petig)</para></listitem>
9254 <listitem><para>Fix for GRANT parameters (Lee Kindness)</para></listitem>
9255 <listitem><para>Fix INITIALLY DEFERRED bug</para></listitem>
9256 <listitem><para>Various bug fixes (Michael, Christof Petig)</para></listitem>
9257 <listitem><para>Auto allocation for indicator variable arrays (int *ind_p=NULL)</para></listitem>
9258 <listitem><para>Auto allocation for string arrays (char **foo_pp=NULL)</para></listitem>
9259 <listitem><para>ECPGfree_auto_mem fixed</para></listitem>
9260 <listitem><para>All function names with external linkage are now prefixed by ECPG</para></listitem>
9261 <listitem><para>Fixes for arrays of structures (Michael)</para></listitem>
9266 <title>Misc. Interfaces</title>
9268 <listitem><para>Python fix fetchone() (Gerhard Haring)</para></listitem>
9269 <listitem><para>Use UTF, Unicode in Tcl where appropriate (Vsevolod Lobko, Reinhard Max)</para></listitem>
9270 <listitem><para>Add Tcl COPY TO/FROM (ljb)</para></listitem>
9271 <listitem><para>Prevent output of default index op class in <application>pg_dump</> (Tom)</para></listitem>
9272 <listitem><para>Fix libpgeasy memory leak (Bruce)</para></listitem>
9277 <title>Build and Install</title>
9279 <listitem><para>Configure, dynamic loader, and shared library fixes (Peter E)</para></listitem>
9280 <listitem><para>Fixes in QNX 4 port (Bernd Tegge)</para></listitem>
9281 <listitem><para>Fixes in Cygwin and Windows ports (Jason Tishler, Gerhard Haring, Dmitry Yurtaev, Darko Prenosil, Mikhail Terekhov)</para></listitem>
9282 <listitem><para>Fix for Windows socket communication failures (Magnus, Mikhail Terekhov)</para></listitem>
9283 <listitem><para>Hurd compile fix (Oliver Elphick)</para></listitem>
9284 <listitem><para>BeOS fixes (Cyril Velter)</para></listitem>
9285 <listitem><para>Remove configure --enable-unicode-conversion, now enabled by multibyte (Tatsuo)</para></listitem>
9286 <listitem><para>AIX fixes (Tatsuo, Andreas)</para></listitem>
9287 <listitem><para>Fix parallel make (Peter E)</para></listitem>
9288 <listitem><para>Install SQL language manual pages into OS-specific directories (Peter E)</para></listitem>
9289 <listitem><para>Rename config.h to pg_config.h (Peter E)</para></listitem>
9290 <listitem><para>Reorganize installation layout of header files (Peter E)</para></listitem>
9295 <title>Source Code</title>
9297 <listitem><para>Remove SEP_CHAR (Bruce)</para></listitem>
9298 <listitem><para>New GUC hooks (Tom)</para></listitem>
9299 <listitem><para>Merge GUC and command line handling (Marko Kreen)</para></listitem>
9300 <listitem><para>Remove EXTEND INDEX (Martijn van Oosterhout, Tom)</para></listitem>
9301 <listitem><para>New pgjindent utility to indent java code (Bruce)</para></listitem>
9302 <listitem><para>Remove define of true/false when compiling under C++ (Leandro Fanzone, Tom)</para></listitem>
9303 <listitem><para>pgindent fixes (Bruce, Tom)</para></listitem>
9304 <listitem><para>Replace strcasecmp() with strcmp() where appropriate (Peter E)</para></listitem>
9305 <listitem><para>Dynahash portability improvements (Tom)</para></listitem>
9306 <listitem><para>Add 'volatile' usage in spinlock structures</para></listitem>
9307 <listitem><para>Improve signal handling logic (Tom)</para></listitem>
9312 <title>Contrib</title>
9314 <listitem><para>New contrib/rtree_gist (Oleg Bartunov, Teodor Sigaev)</para></listitem>
9315 <listitem><para>New contrib/tsearch full-text indexing (Oleg, Teodor Sigaev)</para></listitem>
9316 <listitem><para>Add contrib/dblink for remote database access (Joe Conway)</para></listitem>
9317 <listitem><para>contrib/ora2pg Oracle conversion utility (Gilles Darold)</para></listitem>
9318 <listitem><para>contrib/xml XML conversion utility (John Gray)</para></listitem>
9319 <listitem><para>contrib/fulltextindex fixes (Christopher Kings-Lynne)</para></listitem>
9320 <listitem><para>New contrib/fuzzystrmatch with levenshtein and metaphone, soundex merged (Joe Conway)</para></listitem>
9321 <listitem><para>Add contrib/intarray boolean queries, binary search, fixes (Oleg Bartunov)</para></listitem>
9322 <listitem><para>New pg_upgrade utility (Bruce)</para></listitem>
9323 <listitem><para>Add new pg_resetxlog options (Bruce, Tom)</para></listitem>
9330 <sect1 id="release-7-1-3">
9331 <title>Release 7.1.3</title>
9334 <title>Release date</title>
9335 <simpara>2001-08-15</simpara>
9339 <title>Migration to version 7.1.3</title>
9342 A dump/restore is <emphasis>not</emphasis> required for those running
9348 <title>Changes</title>
9352 Remove unused WAL segements of large transactions (Tom)
9353 Multiaction rule fix (Tom)
9354 PL/pgSQL memory allocation fix (Jan)
9355 VACUUM buffer fix (Tom)
9356 Regression test fixes (Tom)
9357 pg_dump fixes for GRANT/REVOKE/comments on views, user-defined types (Tom)
9358 Fix subselects with DISTINCT ON or LIMIT (Tom)
9360 Disable COPY TO/FROM a view (Tom)
9361 Cygwin build (Jason Tishler)
9368 <sect1 id="release-7-1-2">
9369 <title>Release 7.1.2</title>
9372 <title>Release date</title>
9373 <simpara>2001-05-11</simpara>
9377 This has one fix from 7.1.1.
9382 <title>Migration to version 7.1.2</title>
9385 A dump/restore is <emphasis>not</emphasis> required for those running
9391 <title>Changes</title>
9395 Fix PL/pgSQL SELECTs when returning no rows
9396 Fix for psql backslash core dump
9397 Referential integrity privilege fix
9406 <sect1 id="release-7-1-1">
9407 <title>Release 7.1.1</title>
9410 <title>Release date</title>
9411 <simpara>2001-05-05</simpara>
9415 This has a variety of fixes from 7.1.
9420 <title>Migration to version 7.1.1</title>
9423 A dump/restore is <emphasis>not</emphasis> required for those running
9429 <title>Changes</title>
9433 Fix for numeric MODULO operator (Tom)
9434 pg_dump fixes (Philip)
9435 pg_dump can dump 7.0 databases (Philip)
9436 readline 4.2 fixes (Peter E)
9438 AIX, MSWIN, VAX, N32K fixes (Tom)
9439 Multibytes fixes (Tom)
9440 Unicode fixes (Tatsuo)
9441 Optimizer improvements (Tom)
9442 Fix for whole rows in functions (Tom)
9443 Fix for pg_ctl and option strings with spaces (Peter E)
9444 ODBC fixes (Hiroshi)
9445 EXTRACT can now take string argument (Thomas)
9446 Python fixes (Darcy)
9453 <sect1 id="release-7-1">
9454 <title>Release 7.1</title>
9457 <title>Release date</title>
9458 <simpara>2001-04-13</simpara>
9462 This release focuses on removing limitations that have existed in the
9463 <productname>PostgreSQL</productname> code for many years.
9467 Major changes in this release:
9473 Write-ahead Log (WAL)
9477 To maintain database consistency in case of an operating system crash,
9478 previous releases of <productname>PostgreSQL</productname> have forced
9479 all data modifications to disk before each transaction commit. With
9480 WAL, only one log file must be flushed to disk, greatly improving
9481 performance. If you have been using -F in previous releases to
9482 disable disk flushes, you may want to consider discontinuing its use.
9493 TOAST - Previous releases had a compiled-in row length limit,
9494 typically 8k - 32k. This limit made storage of long text fields
9495 difficult. With TOAST, long rows of any length can be stored with good
9507 We now support outer joins. The UNION/NOT IN
9508 workaround for outer joins is no longer required. We use the SQL92
9520 The previous C function manager did not
9521 handle null values properly, nor did it support 64-bit <acronym>CPU</acronym>'s (Alpha). The new
9522 function manager does. You can continue using your old custom
9523 functions, but you may want to rewrite them in the future to use the new
9524 function manager call interface.
9535 A large number of complex queries that were
9536 unsupported in previous releases now work. Many combinations of views,
9537 aggregates, UNION, LIMIT, cursors, subqueries, and inherited tables
9538 now work properly. Inherited tables are now accessed by default.
9539 Subqueries in FROM are now supported.
9547 <title>Migration to version 7.1</title>
9550 A dump/restore using pg_dump is required for those wishing to migrate
9551 data from any previous release.
9556 <title>Changes</title>
9562 Many multibyte/Unicode/locale fixes (Tatsuo and others)
9563 More reliable ALTER TABLE RENAME (Tom)
9564 Kerberos V fixes (David Wragg)
9565 Fix for INSERT INTO...SELECT where targetlist has subqueries (Tom)
9566 Prompt username/password on standard error (Bruce)
9567 Large objects inv_read/inv_write fixes (Tom)
9568 Fixes for to_char(), to_date(), to_ascii(), and to_timestamp() (Karel,
9570 Prevent query expressions from leaking memory (Tom)
9571 Allow UPDATE of arrays elements (Tom)
9572 Wake up lock waiters during cancel (Hiroshi)
9573 Fix rare cursor crash when using hash join (Tom)
9574 Fix for DROP TABLE/INDEX in rolled-back transaction (Hiroshi)
9575 Fix psql crash from \l+ if MULTIBYTE enabled (Peter E)
9576 Fix truncation of rule names during CREATE VIEW (Ross Reedstrom)
9577 Fix PL/perl (Alex Kapranoff)
9578 Disallow LOCK on views (Mark Hollomon)
9579 Disallow INSERT/UPDATE/DELETE on views (Mark Hollomon)
9580 Disallow DROP RULE, CREATE INDEX, TRUNCATE on views (Mark Hollomon)
9581 Allow PL/pgSQL accept non-ASCII identifiers (Tatsuo)
9582 Allow views to proper handle GROUP BY, aggregates, DISTINCT (Tom)
9583 Fix rare failure with TRUNCATE command (Tom)
9584 Allow UNION/INTERSECT/EXCEPT to be used with ALL, subqueries, views,
9585 DISTINCT, ORDER BY, SELECT...INTO (Tom)
9586 Fix parser failures during aborted transactions (Tom)
9587 Allow temporary relations to properly clean up indexes (Bruce)
9588 Fix VACUUM problem with moving rows in same page (Tom)
9589 Modify pg_dump to better handle user-defined items in template1 (Philip)
9590 Allow LIMIT in VIEW (Tom)
9591 Require cursor FETCH to honor LIMIT (Tom)
9592 Allow PRIMARY/FOREIGN Key definitions on inherited columns (Stephan)
9593 Allow ORDER BY, LIMIT in subqueries (Tom)
9594 Allow UNION in CREATE RULE (Tom)
9595 Make ALTER/DROP TABLE rollback-able (Vadim, Tom)
9596 Store initdb collation in pg_control so collation cannot be changed (Tom)
9597 Fix INSERT...SELECT with rules (Tom)
9598 Fix FOR UPDATE inside views and subselects (Tom)
9599 Fix OVERLAPS operators conform to SQL92 spec regarding NULLs (Tom)
9600 Fix lpad() and rpad() to handle length less than input string (Tom)
9601 Fix use of NOTIFY in some rules (Tom)
9602 Overhaul btree code (Tom)
9603 Fix NOT NULL use in Pl/pgSQL variables (Tom)
9604 Overhaul GIST code (Oleg)
9605 Fix CLUSTER to preserve constraints and column default (Tom)
9606 Improved deadlock detection handling (Tom)
9607 Allow multiple SERIAL columns in a table (Tom)
9608 Prevent occasional index corruption (Vadim)
9612 Add OUTER JOINs (Tom)
9613 Function manager overhaul (Tom)
9614 Allow ALTER TABLE RENAME on indexes (Tom)
9615 Improve CLUSTER (Tom)
9616 Improve ps status display for more platforms (Peter E, Marc)
9617 Improve CREATE FUNCTION failure message (Ross)
9618 JDBC improvements (Peter, Travis Bauer, Christopher Cain, William Webber,
9620 Grand Unified Configuration scheme/GUC. Many options can now be set in
9621 data/postgresql.conf, postmaster/postgres flags, or SET commands (Peter E)
9622 Improved handling of file descriptor cache (Tom)
9623 New warning code about auto-created table alias entries (Bruce)
9624 Overhaul initdb process (Tom, Peter E)
9625 Overhaul of inherited tables; inherited tables now accessed by default;
9626 new ONLY key word prevents it (Chris Bitmead, Tom)
9627 ODBC cleanups/improvements (Nick Gorham, Stephan Szabo, Zoltan Kovacs,
9629 Allow renaming of temp tables (Tom)
9630 Overhaul memory manager contexts (Tom)
9631 pg_dumpall uses CREATE USER or CREATE GROUP rather using COPY (Peter E)
9632 Overhaul pg_dump (Philip Warner)
9633 Allow pg_hba.conf secondary password file to specify only username (Peter E)
9634 Allow TEMPORARY or TEMP key word when creating temporary tables (Bruce)
9635 New memory leak checker (Karel)
9636 New SET SESSION CHARACTERISTICS (Thomas)
9637 Allow nested block comments (Thomas)
9638 Add WITHOUT TIME ZONE type qualifier (Thomas)
9639 New ALTER TABLE ADD CONSTRAINT (Stephan)
9640 Use NUMERIC accumulators for INTEGER aggregates (Tom)
9641 Overhaul aggregate code (Tom)
9642 New VARIANCE and STDDEV() aggregates
9643 Improve dependency ordering of pg_dump (Philip)
9644 New pg_restore command (Philip)
9645 New pg_dump tar output option (Philip)
9646 New pg_dump of large objects (Philip)
9647 New ESCAPE option to LIKE (Thomas)
9648 New case-insensitive LIKE - ILIKE (Thomas)
9649 Allow functional indexes to use binary-compatible type (Tom)
9650 Allow SQL functions to be used in more contexts (Tom)
9651 New pg_config utility (Peter E)
9652 New PL/pgSQL EXECUTE command which allows dynamic SQL and utility statements
9654 New PL/pgSQL GET DIAGNOSTICS statement for SPI value access (Jan)
9655 New quote_identifiers() and quote_literal() functions (Jan)
9656 New ALTER TABLE table OWNER TO user command (Mark Hollomon)
9657 Allow subselects in FROM, i.e. FROM (SELECT ...) [AS] alias (Tom)
9658 Update PyGreSQL to version 3.1 (D'Arcy)
9659 Store tables as files named by OID (Vadim)
9660 New SQL function setval(seq,val,bool) for use in pg_dump (Philip)
9661 Require DROP VIEW to remove views, no DROP TABLE (Mark)
9662 Allow DROP VIEW view1, view2 (Mark)
9663 Allow multiple objects in DROP INDEX, DROP RULE, and DROP TYPE (Tom)
9664 Allow automatic conversion to/from Unicode (Tatsuo, Eiji)
9665 New /contrib/pgcrypto hashing functions (Marko Kreen)
9666 New pg_dumpall --globals-only option (Peter E)
9667 New CHECKPOINT command for WAL which creates new WAL log file (Vadim)
9668 New AT TIME ZONE syntax (Thomas)
9669 Allow location of Unix domain socket to be configurable (David J. MacKenzie)
9670 Allow postmaster to listen on a specific IP address (David J. MacKenzie)
9671 Allow socket path name to be specified in hostname by using leading slash
9672 (David J. MacKenzie)
9673 Allow CREATE DATABASE to specify template database (Tom)
9674 New utility to convert MySQL schema dumps to SQL92 and PostgreSQL (Thomas)
9675 New /contrib/rserv replication toolkit (Vadim)
9676 New file format for COPY BINARY (Tom)
9677 New /contrib/oid2name to map numeric files to table names (B Palmer)
9678 New "idle in transaction" ps status message (Marc)
9679 Update to pgaccess 0.98.7 (Constantin Teodorescu)
9680 pg_ctl now defaults to -w (wait) on shutdown, new -l (log) option
9681 Add rudimentary dependency checking to pg_dump (Philip)
9685 Fix INET/CIDR type ordering and add new functions (Tom)
9686 Make OID behave as an unsigned type (Tom)
9687 Allow BIGINT as synonym for INT8 (Peter E)
9688 New int2 and int8 comparison operators (Tom)
9689 New BIT and BIT VARYING types (Adriaan Joubert, Tom, Peter E)
9690 CHAR() no longer faster than VARCHAR() because of TOAST (Tom)
9691 New GIST seg/cube examples (Gene Selkov)
9692 Improved round(numeric) handling (Tom)
9693 Fix CIDR output formatting (Tom)
9694 New CIDR abbrev() function (Tom)
9698 Write-Ahead Log (WAL) to provide crash recovery with less performance
9700 ANALYZE stage of VACUUM no longer exclusively locks table (Bruce)
9701 Reduced file seeks (Denis Perchine)
9702 Improve BTREE code for duplicate keys (Tom)
9703 Store all large objects in a single table (Denis Perchine, Tom)
9704 Improve memory allocation performance (Karel, Tom)
9708 New function manager call conventions (Tom)
9709 SGI portability fixes (David Kaelbling)
9710 New configure --enable-syslog option (Peter E)
9711 New BSDI README (Bruce)
9712 configure script moved to top level, not /src (Peter E)
9713 Makefile/configuration/compilation overhaul (Peter E)
9714 New configure --with-python option (Peter E)
9715 Solaris cleanups (Peter E)
9716 Overhaul /contrib Makefiles (Karel)
9717 New OpenSSL configuration option (Magnus, Peter E)
9719 QNX fixes (Maurizio)
9720 New heap_open(), heap_openr() API (Tom)
9721 Remove colon and semi-colon operators (Thomas)
9722 New pg_class.relkind value for views (Mark Hollomon)
9723 Rename ichar() to chr() (Karel)
9724 New documentation for btrim(), ascii(), chr(), repeat() (Karel)
9725 Fixes for NT/Cygwin (Pete Forman)
9726 AIX port fixes (Andreas)
9727 New BeOS port (David Reid, Cyril Velter)
9728 Add proofreader's changes to docs (Addison-Wesley, Bruce)
9729 New Alpha spinlock code (Adriaan Joubert, Compaq)
9730 UnixWare port overhaul (Peter E)
9731 New Darwin/MacOS X port (Peter Bierman, Bruce Hartzler)
9732 New FreeBSD Alpha port (Alfred)
9733 Overhaul shared memory segments (Tom)
9734 Add IBM S/390 support (Neale Ferguson)
9735 Moved macmanuf to /contrib (Larry Rosenman)
9736 Syslog improvements (Larry Rosenman)
9737 New template0 database that contains no user additions (Tom)
9738 New /contrib/cube and /contrib/seg GIST sample code (Gene Selkov)
9739 Allow NetBSD's libedit instead of readline (Peter)
9740 Improved assembly language source code format (Bruce)
9741 New contrib/pg_logger
9742 New --template option to createdb
9743 New contrib/pg_control utility (Oliver)
9744 New FreeBSD tools ipc_check, start-scripts/freebsd
9751 <sect1 id="release-7-0-3">
9752 <title>Release 7.0.3</title>
9755 <title>Release date</title>
9756 <simpara>2000-11-11</simpara>
9760 This has a variety of fixes from 7.0.2.
9765 <title>Migration to version 7.0.3</title>
9768 A dump/restore is <emphasis>not</emphasis> required for those running
9774 <title>Changes</title>
9779 Large object fix (Tom)
9780 Fix lean in COPY WITH OIDS leak (Tom)
9781 Fix backwards-index-scan (Tom)
9782 Fix SELECT ... FOR UPDATE so it checks for duplicate keys (Hiroshi)
9783 Add --enable-syslog to configure (Marc)
9784 Fix abort transaction at backend exit in rare cases (Tom)
9785 Fix for psql \l+ when multibyte enabled (Tatsuo)
9786 Allow PL/pgSQL to accept non ascii identifiers (Tatsuo)
9787 Make vacuum always flush buffers (Tom)
9788 Fix to allow cancel while waiting for a lock (Hiroshi)
9789 Fix for memory aloocation problem in user authentication code (Tom)
9790 Remove bogus use of int4out() (Tom)
9791 Fixes for multiple subqueries in COALESCE or BETWEEN (Tom)
9792 Fix for failure of triggers on heap open in certain cases (Jeroen van
9794 Fix for erroneous selectivity of not-equals (Tom)
9795 Fix for erroneous use of strcmp() (Tom)
9796 Fix for bug where storage manager accesses items beyond end of file
9798 Fix to include kernel errno message in all smgr elog messages (Tom)
9799 Fix for '.' not in PATH at build time (SL Baur)
9800 Fix for out-of-file-descriptors error (Tom)
9801 Fix to make pg_dump dump 'iscachable' flag for functions (Tom)
9802 Fix for subselect in targetlist of Append node (Tom)
9803 Fix for mergejoin plans (Tom)
9804 Fix TRUNCATE failure on relations with indexes (Tom)
9805 Avoid database-wide restart on write error (Hiroshi)
9806 Fix nodeMaterial to honor chgParam by recomputing its output (Tom)
9807 Fix VACUUM problem with moving chain of update row versions when source
9808 and destination of a row version lie on the same page (Tom)
9809 Fix user.c CommandCounterIncrement (Tom)
9810 Fix for AM/PM boundary problem in to_char() (Karel Zak)
9811 Fix TIME aggregate handling (Tom)
9812 Fix to_char() to avoid coredump on NULL input (Tom)
9814 Fix for inserting/copying longer multibyte strings into char() data
9816 Fix for crash of backend, on abort (Tom)
9823 <sect1 id="release-7-0-2">
9824 <title>Release 7.0.2</title>
9827 <title>Release date</title>
9828 <simpara>2000-06-05</simpara>
9832 This is a repackaging of 7.0.1 with added documentation.
9837 <title>Migration to version 7.0.2</title>
9840 A dump/restore is <emphasis>not</emphasis> required for those running
9846 <title>Changes</title>
9850 Added documentation to tarball.
9857 <sect1 id="release-7-0-1">
9858 <title>Release 7.0.1</title>
9861 <title>Release date</title>
9862 <simpara>2000-06-01</simpara>
9866 This is a cleanup release for 7.0.
9870 <title>Migration to version 7.0.1</title>
9873 A dump/restore is <emphasis>not</emphasis> required for those running
9879 <title>Changes</title>
9883 Fix many CLUSTER failures (Tom)
9884 Allow ALTER TABLE RENAME works on indexes (Tom)
9885 Fix plpgsql to handle datetime->timestamp and timespan->interval (Bruce)
9886 New configure --with-setproctitle switch to use setproctitle() (Marc, Bruce)
9887 Fix the off by one errors in ResultSet from 6.5.3, and more.
9888 jdbc ResultSet fixes (Joseph Shraibman)
9889 optimizer tunings (Tom)
9890 Fix create user for pgaccess
9891 Fix for UNLISTEN failure
9892 IRIX fixes (David Kaelbling)
9893 QNX fixes (Andreas Kardos)
9894 Reduce COPY IN lock level (Tom)
9895 Change libpqeasy to use PQconnectdb() style parameters (Bruce)
9896 Fix pg_dump to handle OID indexes (Tom)
9897 Fix small memory leak (Tom)
9898 Solaris fix for createdb/dropdb (Tatsuo)
9899 Fix for non-blocking connections (Alfred Perlstein)
9900 Fix improper recovery after RENAME TABLE failures (Tom)
9901 Copy pg_ident.conf.sample into /lib directory in install (Bruce)
9902 Add SJIS UDC (NEC selection IBM kanji) support (Eiji Tokuya)
9903 Fix too long syslog message (Tatsuo)
9904 Fix problem with quoted indexes that are too long (Tom)
9905 JDBC ResultSet.getTimestamp() fix (Gregory Krasnow & Floyd Marinescu)
9906 ecpg changes (Michael)
9912 <sect1 id="release-7-0">
9913 <title>Release 7.0</title>
9916 <title>Release date</title>
9917 <simpara>2000-05-08</simpara>
9921 This release contains improvements in many areas, demonstrating
9922 the continued growth of <productname>PostgreSQL</productname>.
9923 There are more improvements and fixes in 7.0 than in any previous
9924 release. The developers have confidence that this is the best
9925 release yet; we do our best to put out only solid releases, and
9926 this one is no exception.
9930 Major changes in this release:
9940 Foreign keys are now implemented, with the exception of PARTIAL MATCH
9941 foreign keys. Many users have been asking for this feature, and we are
9942 pleased to offer it.
9953 Continuing on work started a year ago, the optimizer has been
9954 improved, allowing better query plan selection and faster performance
9955 with less memory usage.
9962 Updated <application>psql</application>
9966 <application>psql</application>, our interactive terminal monitor, has been
9967 updated with a variety of new features. See the <application>psql</application> manual page for details.
9978 SQL92 join syntax is now supported, though only as
9979 <literal>INNER JOIN</> for this release. <literal>JOIN</>,
9980 <literal>NATURAL JOIN</>, <literal>JOIN</>/<literal>USING</>,
9981 and <literal>JOIN</>/<literal>ON</> are available, as are
9982 column correlation names.
9990 <title>Migration to version 7.0</title>
9993 A dump/restore using <application>pg_dump</application>
9994 is required for those wishing to migrate data from any
9995 previous release of <productname>PostgreSQL</productname>.
9996 For those upgrading from 6.5.*, you may instead use
9997 <application>pg_upgrade</application> to upgrade to this
9998 release; however, a full dump/reload installation is always the
9999 most robust method for upgrades.
10003 Interface and compatibility issues to consider for the new
10010 The date/time types <type>datetime</type> and
10011 <type>timespan</type> have been superseded by the
10012 SQL92-defined types <type>timestamp</type> and
10013 <type>interval</type>. Although there has been some effort to
10014 ease the transition by allowing
10015 <productname>PostgreSQL</productname> to recognize
10016 the deprecated type names and translate them to the new type
10017 names, this mechanism may not be completely transparent to
10018 your existing application.
10024 The optimizer has been substantially improved in the area of
10025 query cost estimation. In some cases, this will result in
10026 decreased query times as the optimizer makes a better choice
10027 for the preferred plan. However, in a small number of cases,
10028 usually involving pathological distributions of data, your
10029 query times may go up. If you are dealing with large amounts
10030 of data, you may want to check your queries to verify
10037 The <acronym>JDBC</acronym> and <acronym>ODBC</acronym>
10038 interfaces have been upgraded and extended.
10044 The string function <function>CHAR_LENGTH</function> is now a
10045 native function. Previous versions translated this into a call
10046 to <function>LENGTH</function>, which could result in
10047 ambiguity with other types implementing
10048 <function>LENGTH</function> such as the geometric types.
10055 <title>Changes</title>
10061 Prevent function calls exceeding maximum number of arguments (Tom)
10062 Improve CASE construct (Tom)
10063 Fix SELECT coalesce(f1,0) FROM int4_tbl GROUP BY f1 (Tom)
10064 Fix SELECT sentence.words[0] FROM sentence GROUP BY sentence.words[0] (Tom)
10065 Fix GROUP BY scan bug (Tom)
10066 Improvements in SQL grammar processing (Tom)
10067 Fix for views involved in INSERT ... SELECT ... (Tom)
10068 Fix for SELECT a/2, a/2 FROM test_missing_target GROUP BY a/2 (Tom)
10069 Fix for subselects in INSERT ... SELECT (Tom)
10070 Prevent INSERT ... SELECT ... ORDER BY (Tom)
10071 Fixes for relations greater than 2GB, including vacuum
10072 Improve propagating system table changes to other backends (Tom)
10073 Improve propagating user table changes to other backends (Tom)
10074 Fix handling of temp tables in complex situations (Bruce, Tom)
10075 Allow table locking at table open, improving concurrent reliability (Tom)
10076 Properly quote sequence names in pg_dump (Ross J. Reedstrom)
10077 Prevent DROP DATABASE while others accessing
10078 Prevent any rows from being returned by GROUP BY if no rows processed (Tom)
10079 Fix SELECT COUNT(1) FROM table WHERE ...' if no rows matching WHERE (Tom)
10080 Fix pg_upgrade so it works for MVCC (Tom)
10081 Fix for SELECT ... WHERE x IN (SELECT ... HAVING SUM(x) > 1) (Tom)
10082 Fix for "f1 datetime DEFAULT 'now'" (Tom)
10083 Fix problems with CURRENT_DATE used in DEFAULT (Tom)
10084 Allow comment-only lines, and ;;; lines too. (Tom)
10085 Improve recovery after failed disk writes, disk full (Hiroshi)
10086 Fix cases where table is mentioned in FROM but not joined (Tom)
10087 Allow HAVING clause without aggregate functions (Tom)
10088 Fix for "--" comment and no trailing newline, as seen in perl interface
10089 Improve pg_dump failure error reports (Bruce)
10090 Allow sorts and hashes to exceed 2GB file sizes (Tom)
10091 Fix for pg_dump dumping of inherited rules (Tom)
10092 Fix for NULL handling comparisons (Tom)
10093 Fix inconsistent state caused by failed CREATE/DROP commands (Hiroshi)
10094 Fix for dbname with dash
10095 Prevent DROP INDEX from interfering with other backends (Tom)
10096 Fix file descriptor leak in verify_password()
10097 Fix for "Unable to identify an operator =$" problem
10098 Fix ODBC so no segfault if CommLog and Debug enabled (Dirk Niggemann)
10099 Fix for recursive exit call (Massimo)
10100 Fix for extra-long timezones (Jeroen van Vianen)
10101 Make pg_dump preserve primary key information (Peter E)
10102 Prevent databases with single quotes (Peter E)
10103 Prevent DROP DATABASE inside transaction (Peter E)
10104 ecpg memory leak fixes (Stephen Birch)
10105 Fix for SELECT null::text, SELECT int4fac(null) and SELECT 2 + (null) (Tom)
10106 Y2K timestamp fix (Massimo)
10107 Fix for VACUUM 'HEAP_MOVED_IN was not expected' errors (Tom)
10108 Fix for views with tables/columns containing spaces (Tom)
10109 Prevent privileges on indexes (Peter E)
10110 Fix for spinlock stuck problem when error is generated (Hiroshi)
10111 Fix ipcclean on Linux
10112 Fix handling of NULL constraint conditions (Tom)
10113 Fix memory leak in odbc driver (Nick Gorham)
10114 Fix for privilege check on UNION tables (Tom)
10115 Fix to allow SELECT 'a' LIKE 'a' (Tom)
10116 Fix for SELECT 1 + NULL (Tom)
10118 Fix log() on numeric type (Tom)
10119 Deprecate ':' and ';' operators
10120 Allow vacuum of temporary tables
10121 Disallow inherited columns with the same name as new columns
10122 Recover or force failure when disk space is exhausted (Hiroshi)
10123 Fix INSERT INTO ... SELECT with AS columns matching result columns
10124 Fix INSERT ... SELECT ... GROUP BY groups by target columns not source columns (Tom)
10125 Fix CREATE TABLE test (a char(5) DEFAULT text '', b int4) with INSERT (Tom)
10126 Fix UNION with LIMIT
10127 Fix CREATE TABLE x AS SELECT 1 UNION SELECT 2
10128 Fix CREATE TABLE test(col char(2) DEFAULT user)
10129 Fix mismatched types in CREATE TABLE ... DEFAULT
10130 Fix SELECT * FROM pg_class where oid in (0,-1)
10131 Fix SELECT COUNT('asdf') FROM pg_class WHERE oid=12
10132 Prevent user who can create databases can modifying pg_database table (Peter E)
10133 Fix btree to give a useful elog when key > 1/2 (page - overhead) (Tom)
10134 Fix INSERT of 0.0 into DECIMAL(4,4) field (Tom)
10138 New CLI interface include file sqlcli.h, based on SQL3/SQL98
10139 Remove all limits on query length, row length limit still exists (Tom)
10140 Update jdbc protocol to 2.0 (Jens Glaser <email>jens@jens.de</email>)
10141 Add TRUNCATE command to quickly truncate relation (Mike Mascari)
10142 Fix to give super user and createdb user proper update catalog rights (Peter E)
10143 Allow ecpg bool variables to have NULL values (Christof)
10144 Issue ecpg error if NULL value for variable with no NULL indicator (Christof)
10145 Allow ^C to cancel COPY command (Massimo)
10146 Add SET FSYNC and SHOW PG_OPTIONS commands(Massimo)
10147 Function name overloading for dynamically-loaded C functions (Frankpitt)
10148 Add CmdTuples() to libpq++(Vince)
10149 New CREATE CONSTRAINT TRIGGER and SET CONSTRAINTS commands(Jan)
10150 Allow CREATE FUNCTION/WITH clause to be used for all language types
10151 configure --enable-debug adds -g (Peter E)
10152 configure --disable-debug removes -g (Peter E)
10153 Allow more complex default expressions (Tom)
10154 First real FOREIGN KEY constraint trigger functionality (Jan)
10155 Add FOREIGN KEY ... MATCH FULL ... ON DELETE CASCADE (Jan)
10156 Add FOREIGN KEY ... MATCH <unspecified> referential actions (Don Baccus)
10157 Allow WHERE restriction on ctid (physical heap location) (Hiroshi)
10158 Move pginterface from contrib to interface directory, rename to pgeasy (Bruce)
10159 Change pgeasy connectdb() parameter ordering (Bruce)
10160 Require SELECT DISTINCT target list to have all ORDER BY columns (Tom)
10161 Add Oracle's COMMENT ON command (Mike Mascari <email>mascarim@yahoo.com</email>)
10162 libpq's PQsetNoticeProcessor function now returns previous hook(Peter E)
10163 Prevent PQsetNoticeProcessor from being set to NULL (Peter E)
10164 Make USING in COPY optional (Bruce)
10165 Allow subselects in the target list (Tom)
10166 Allow subselects on the left side of comparison operators (Tom)
10167 New parallel regression test (Jan)
10168 Change backend-side COPY to write files with permissions 644 not 666 (Tom)
10169 Force permissions on PGDATA directory to be secure, even if it exists (Tom)
10170 Added psql LASTOID variable to return last inserted oid (Peter E)
10171 Allow concurrent vacuum and remove pg_vlock vacuum lock file (Tom)
10172 Add privilege check for vacuum (Peter E)
10173 New libpq functions to allow asynchronous connections: PQconnectStart(),
10174 PQconnectPoll(), PQresetStart(), PQresetPoll(), PQsetenvStart(),
10175 PQsetenvPoll(), PQsetenvAbort (Ewan Mellor)
10176 New libpq PQsetenv() function (Ewan Mellor)
10177 create/alter user extension (Peter E)
10178 New postmaster.pid and postmaster.opts under $PGDATA (Tatsuo)
10179 New scripts for create/drop user/db (Peter E)
10180 Major psql overhaul (Peter E)
10181 Add const to libpq interface (Peter E)
10182 New libpq function PQoidValue (Peter E)
10183 Show specific non-aggregate causing problem with GROUP BY (Tom)
10184 Make changes to pg_shadow recreate pg_pwd file (Peter E)
10185 Add aggregate(DISTINCT ...) (Tom)
10186 Allow flag to control COPY input/output of NULLs (Peter E)
10187 Make postgres user have a password by default (Peter E)
10188 Add CREATE/ALTER/DROP GROUP (Peter E)
10189 All administration scripts now support --long options (Peter E, Karel)
10190 Vacuumdb script now supports --all option (Peter E)
10191 ecpg new portable FETCH syntax
10192 Add ecpg EXEC SQL IFDEF, EXEC SQL IFNDEF, EXEC SQL ELSE, EXEC SQL ELIF
10193 and EXEC SQL ENDIF directives
10194 Add pg_ctl script to control backend start-up (Tatsuo)
10195 Add postmaster.opts.default file to store start-up flags (Tatsuo)
10196 Allow --with-mb=SQL_ASCII
10197 Increase maximum number of index keys to 16 (Bruce)
10198 Increase maximum number of function arguments to 16 (Bruce)
10199 Allow configuration of maximum number of index keys and arguments (Bruce)
10200 Allow unprivileged users to change their passwords (Peter E)
10201 Password authentication enabled; required for new users (Peter E)
10202 Disallow dropping a user who owns a database (Peter E)
10203 Change initdb option --with-mb to --enable-multibyte
10204 Add option for initdb to prompts for superuser password (Peter E)
10205 Allow complex type casts like col::numeric(9,2) and col::int2::float8 (Tom)
10206 Updated user interfaces on initdb, initlocation, pg_dump, ipcclean (Peter E)
10207 New pg_char_to_encoding() and pg_encoding_to_char() functions (Tatsuo)
10208 libpq non-blocking mode (Alfred Perlstein)
10209 Improve conversion of types in casts that don't specify a length
10210 New plperl internal programming language (Mark Hollomon)
10211 Allow COPY IN to read file that do not end with a newline (Tom)
10212 Indicate when long identifiers are truncated (Tom)
10213 Allow aggregates to use type equivalency (Peter E)
10214 Add Oracle's to_char(), to_date(), to_datetime(), to_timestamp(), to_number()
10215 conversion functions (Karel Zak <zakkr@zf.jcu.cz>)
10216 Add SELECT DISTINCT ON (expr [, expr ...]) targetlist ... (Tom)
10217 Check to be sure ORDER BY is compatible with the DISTINCT operation (Tom)
10218 Add NUMERIC and int8 types to ODBC
10219 Improve EXPLAIN results for Append, Group, Agg, Unique (Tom)
10220 Add ALTER TABLE ... ADD FOREIGN KEY (Stephan Szabo)
10221 Allow SELECT .. FOR UPDATE in PL/pgSQL (Hiroshi)
10222 Enable backward sequential scan even after reaching EOF (Hiroshi)
10223 Add btree indexing of boolean values, >= and <= (Don Baccus)
10224 Print current line number when COPY FROM fails (Massimo)
10225 Recognize POSIX time zone e.g. "PST+8" and "GMT-8" (Thomas)
10226 Add DEC as synonym for DECIMAL (Thomas)
10227 Add SESSION_USER as SQL92 key word, same as CURRENT_USER (Thomas)
10228 Implement SQL92 column aliases (aka correlation names) (Thomas)
10229 Implement SQL92 join syntax (Thomas)
10230 Make INTERVAL reserved word allowed as a column identifier (Thomas)
10231 Implement REINDEX command (Hiroshi)
10232 Accept ALL in aggregate function SUM(ALL col) (Tom)
10233 Prevent GROUP BY from using column aliases (Tom)
10234 New psql \encoding option (Tatsuo)
10235 Allow PQrequestCancel() to terminate when in waiting-for-lock state (Hiroshi)
10236 Allow negation of a negative number in all cases
10237 Add ecpg descriptors (Christof, Michael)
10238 Allow CREATE VIEW v AS SELECT f1::char(8) FROM tbl
10239 Allow casts with length, like foo::char(8)
10240 New libpq functions PQsetClientEncoding(), PQclientEncoding() (Tatsuo)
10241 Add support for SJIS user defined characters (Tatsuo)
10242 Larger views/rules supported
10243 Make libpq's PQconndefaults() thread-safe (Tom)
10244 Disable // as comment to be ANSI conforming, should use -- (Tom)
10245 Allow column aliases on views CREATE VIEW name (collist)
10246 Fixes for views with subqueries (Tom)
10247 Allow UPDATE table SET fld = (SELECT ...) (Tom)
10248 SET command options no longer require quotes
10249 Update pgaccess to 0.98.6
10250 New SET SEED command
10251 New pg_options.sample file
10252 New SET FSYNC command (Massimo)
10253 Allow pg_descriptions when creating tables
10254 Allow pg_descriptions when creating types, columns, and functions
10255 Allow psql \copy to allow delimiters (Peter E)
10256 Allow psql to print nulls as distinct from "" [null] (Peter E)
10260 Many array fixes (Tom)
10261 Allow bare column names to be subscripted as arrays (Tom)
10262 Improve type casting of int and float constants (Tom)
10263 Cleanups for int8 inputs, range checking, and type conversion (Tom)
10264 Fix for SELECT timespan('21:11:26'::time) (Tom)
10265 netmask('x.x.x.x/0') is 255.255.255.255 instead of 0.0.0.0 (Oleg Sharoiko)
10266 Add btree index on NUMERIC (Jan)
10267 Perl fix for large objects containing NUL characters (Douglas Thomson)
10268 ODBC fix for for large objects (free)
10269 Fix indexing of cidr data type
10270 Fix for Ethernet MAC addresses (macaddr type) comparisons
10271 Fix for date/time types when overflows happened in computations (Tom)
10272 Allow array on int8 (Peter E)
10273 Fix for rounding/overflow of NUMERIC type, like NUMERIC(4,4) (Tom)
10274 Allow NUMERIC arrays
10275 Fix bugs in NUMERIC ceil() and floor() functions (Tom)
10276 Make char_length()/octet_length including trailing blanks (Tom)
10277 Made abstime/reltime use int4 instead of time_t (Peter E)
10278 New lztext data type for compressed text fields
10279 Revise code to handle coercion of int and float constants (Tom)
10280 Start at new code to implement a BIT and BIT VARYING type (Adriaan Joubert)
10281 NUMERIC now accepts scientific notation (Tom)
10282 NUMERIC to int4 rounds (Tom)
10283 Convert float4/8 to NUMERIC properly (Tom)
10284 Allow type conversion with NUMERIC (Thomas)
10285 Make ISO date style (2000-02-16 09:33) the default (Thomas)
10286 Add NATIONAL CHAR [ VARYING ] (Thomas)
10287 Allow NUMERIC round and trunc to accept negative scales (Tom)
10288 New TIME WITH TIME ZONE type (Thomas)
10289 Add MAX()/MIN() on time type (Thomas)
10290 Add abs(), mod(), fac() for int8 (Thomas)
10291 Rename functions to round(), sqrt(), cbrt(), pow() for float8 (Thomas)
10292 Add transcendental math functions (e.g. sin(), acos()) for float8 (Thomas)
10293 Add exp() and ln() for NUMERIC type
10294 Rename NUMERIC power() to pow() (Thomas)
10295 Improved TRANSLATE() function (Edwin Ramirez, Tom)
10296 Allow X=-Y operators (Tom)
10297 Allow SELECT float8(COUNT(*))/(SELECT COUNT(*) FROM t) FROM t GROUP BY f1; (Tom)
10298 Allow LOCALE to use indexes in regular expression searches (Tom)
10299 Allow creation of functional indexes to use default types
10303 Prevent exponential space consumption with many AND's and OR's (Tom)
10304 Collect attribute selectivity values for system columns (Tom)
10305 Reduce memory usage of aggregates (Tom)
10306 Fix for LIKE optimization to use indexes with multibyte encodings (Tom)
10307 Fix r-tree index optimizer selectivity (Thomas)
10308 Improve optimizer selectivity computations and functions (Tom)
10309 Optimize btree searching for cases where many equal keys exist (Tom)
10310 Enable fast LIKE index processing only if index present (Tom)
10311 Re-use free space on index pages with duplicates (Tom)
10312 Improve hash join processing (Tom)
10313 Prevent descending sort if result is already sorted(Hiroshi)
10314 Allow commuting of index scan query qualifications (Tom)
10315 Prefer index scans in cases where ORDER BY/GROUP BY is required (Tom)
10316 Allocate large memory requests in fix-sized chunks for performance (Tom)
10317 Fix vacuum's performance by reducing memory allocation requests (Tom)
10318 Implement constant-expression simplification (Bernard Frankpitt, Tom)
10319 Use secondary columns to be used to determine start of index scan (Hiroshi)
10320 Prevent quadruple use of disk space when doing internal sorting (Tom)
10321 Faster sorting by calling fewer functions (Tom)
10322 Create system indexes to match all system caches (Bruce, Hiroshi)
10323 Make system caches use system indexes (Bruce)
10324 Make all system indexes unique (Bruce)
10325 Improve pg_statistics management for VACUUM speed improvement (Tom)
10326 Flush backend cache less frequently (Tom, Hiroshi)
10327 COPY now reuses previous memory allocation, improving performance (Tom)
10328 Improve optimization cost estimation (Tom)
10329 Improve optimizer estimate of range queries x > lowbound AND x < highbound (Tom)
10330 Use DNF instead of CNF where appropriate (Tom, Taral)
10331 Further cleanup for OR-of-AND WHERE-clauses (Tom)
10332 Make use of index in OR clauses (x = 1 AND y = 2) OR (x = 2 AND y = 4) (Tom)
10333 Smarter optimizer computations for random index page access (Tom)
10334 New SET variable to control optimizer costs (Tom)
10335 Optimizer queries based on LIMIT, OFFSET, and EXISTS qualifications (Tom)
10336 Reduce optimizer internal housekeeping of join paths for speedup (Tom)
10337 Major subquery speedup (Tom)
10338 Fewer fsync writes when fsync is not disabled (Tom)
10339 Improved LIKE optimizer estimates (Tom)
10340 Prevent fsync in SELECT-only queries (Vadim)
10341 Make index creation use psort code, because it is now faster (Tom)
10342 Allow creation of sort temp tables > 1 Gig
10344 Source Tree Changes
10345 -------------------
10346 Fix for linux PPC compile
10347 New generic expression-tree-walker subroutine (Tom)
10348 Change form() to varargform() to prevent portability problems
10349 Improved range checking for large integers on Alphas
10350 Clean up #include in /include directory (Bruce)
10351 Add scripts for checking includes (Bruce)
10352 Remove un-needed #include's from *.c files (Bruce)
10353 Change #include's to use <> and "" as appropriate (Bruce)
10354 Enable Windows compilation of libpq
10355 Alpha spinlock fix from Uncle George <email>gatgul@voicenet.com</email>
10356 Overhaul of optimizer data structures (Tom)
10357 Fix to cygipc library (Yutaka Tanida)
10358 Allow pgsql to work on newer Cygwin snapshots (Dan)
10359 New catalog version number (Tom)
10361 Rename heap_replace to heap_update
10362 Update for QNX (Dr. Andreas Kardos)
10363 New platform-specific regression handling (Tom)
10364 Rename oid8 -> oidvector and int28 -> int2vector (Bruce)
10365 Included all yacc and lex files into the distribution (Peter E.)
10366 Remove lextest, no longer needed (Peter E)
10367 Fix for libpq and psql on Windows (Magnus)
10368 Internally change datetime and timespan into timestamp and interval (Thomas)
10369 Fix for plpgsql on BSD/OS
10370 Add SQL_ASCII test case to the regression test (Tatsuo)
10371 configure --with-mb now deprecated (Tatsuo)
10373 NetBSD fixes (Johnny C. Lam <email>lamj@stat.cmu.edu</email>)
10374 Fixes for Alpha compiles
10375 New multibyte encodings
10381 <sect1 id="release-6-5-3">
10382 <title>Release 6.5.3</title>
10385 <title>Release date</title>
10386 <simpara>1999-10-13</simpara>
10390 This is basically a cleanup release for 6.5.2. We have added a new
10391 <application>PgAccess</> that was missing in 6.5.2, and installed an NT-specific fix.
10396 <title>Migration to version 6.5.3</title>
10399 A dump/restore is <emphasis>not</emphasis> required for those running
10404 <title>Changes</title>
10408 Updated version of pgaccess 0.98
10410 Fix dumping rules on inherited tables
10417 <sect1 id="release-6-5-2">
10418 <title>Release 6.5.2</title>
10421 <title>Release date</title>
10422 <simpara>1999-09-15</simpara>
10426 This is basically a cleanup release for 6.5.1. We have fixed a variety of
10427 problems reported by 6.5.1 users.
10432 <title>Migration to version 6.5.2</title>
10435 A dump/restore is <emphasis>not</emphasis> required for those running
10441 <title>Changes</title>
10445 subselect+CASE fixes(Tom)
10446 Add SHLIB_LINK setting for solaris_i386 and solaris_sparc ports(Daren Sefcik)
10447 Fixes for CASE in WHERE join clauses(Tom)
10448 Fix BTScan abort(Tom)
10449 Repair the check for redundant UNIQUE and PRIMARY KEY indexes(Thomas)
10450 Improve it so that it checks for multicolumn constraints(Thomas)
10451 Fix for Windows making problem with MB enabled(Hiroki Kataoka)
10452 Allow BSD yacc and bison to compile pl code(Bruce)
10453 Fix SET NAMES working
10455 Fix vacuum's memory consumption(Hiroshi,Tatsuo)
10456 Reduce the total memory consumption of vacuum(Tom)
10457 Fix for timestamp(datetime)
10458 Rule deparsing bugfixes(Tom)
10459 Fix quoting problems in mkMakefile.tcldefs.sh.in and mkMakefile.tkdefs.sh.in(Tom)
10460 This is to re-use space on index pages freed by vacuum(Vadim)
10461 document -x for pg_dump(Bruce)
10462 Fix for unary operators in rule deparser(Tom)
10463 Comment out FileUnlink of excess segments during mdtruncate()(Tom)
10464 IRIX linking fix from Yu Cao >yucao@falcon.kla-tencor.com<
10465 Repair logic error in LIKE: should not return LIKE_ABORT
10466 when reach end of pattern before end of text(Tom)
10467 Repair incorrect cleanup of heap memory allocation during transaction abort(Tom)
10468 Updated version of pgaccess 0.98
10474 <sect1 id="release-6-5-1">
10475 <title>Release 6.5.1</title>
10478 <title>Release date</title>
10479 <simpara>1999-07-15</simpara>
10483 This is basically a cleanup release for 6.5. We have fixed a variety of
10484 problems reported by 6.5 users.
10488 <title>Migration to version 6.5.1</title>
10491 A dump/restore is <emphasis>not</emphasis> required for those running
10497 <title>Changes</title>
10502 Portability fixes for linux_ppc, IRIX, linux_alpha, OpenBSD, alpha
10503 Remove QUERY_LIMIT, use SELECT...LIMIT
10504 Fix for EXPLAIN on inheritance(Tom)
10505 Patch to allow vacuum on multisegment tables(Hiroshi)
10506 R-Tree optimizer selectivity fix(Tom)
10507 ACL file descriptor leak fix(Atsushi Ogawa)
10508 New expresssion subtree code(Tom)
10509 Avoid disk writes for read-only transactions(Vadim)
10510 Fix for removal of temp tables if last transaction was aborted(Bruce)
10511 Fix to prevent too large row from being created(Bruce)
10513 Allow port numbers 32k - 64k(Bruce)
10514 Add ^ precidence(Bruce)
10515 Rename sort files called pg_temp to pg_sorttemp(Bruce)
10516 Fix for microseconds in time values(Tom)
10517 Tutorial source cleanup
10518 New linux_m68k port
10519 Fix for sorting of NULL's in some cases(Tom)
10520 Shared library dependencies fixed (Tom)
10521 Fixed glitches affecting GROUP BY in subselects(Tom)
10522 Fix some compiler warnings (Tomoaki Nishiyama)
10523 Add Win1250 (Czech) support (Pavel Behal)
10529 <sect1 id="release-6-5">
10530 <title>Release 6.5</title>
10533 <title>Release date</title>
10534 <simpara>1999-06-09</simpara>
10538 This release marks a major step in the development team's mastery of the source
10539 code we inherited from Berkeley. You will see we are now easily adding
10540 major features, thanks to the increasing size and experience of our
10541 world-wide development team.
10545 Here is a brief summary of the more notable changes:
10550 Multiversion concurrency control(MVCC)
10554 This removes our old table-level locking, and replaces it with
10555 a locking system that is superior to most commercial database
10556 systems. In a traditional system, each row that is modified
10557 is locked until committed, preventing reads by other users.
10558 MVCC uses the natural multiversion nature of
10559 <productname>PostgreSQL</productname> to allow readers to
10560 continue reading consistent data during writer activity.
10561 Writers continue to use the compact pg_log transaction system.
10562 This is all performed without having to allocate a lock for
10563 every row like traditional database systems. So, basically,
10564 we no longer are restricted by simple table-level locking; we
10565 have something better than row-level locking.
10572 Hot backups from <application>pg_dump</application>
10576 <application>pg_dump</application> takes advantage of the new
10577 MVCC features to give a consistent database dump/backup while
10578 the database stays online and available for queries.
10589 We now have a true numeric data type, with
10590 user-specified precision.
10601 Temporary tables are guaranteed to have unique names
10602 within a database session, and are destroyed on session exit.
10613 We now have CASE, INTERSECT, and EXCEPT statement
10614 support. We have new LIMIT/OFFSET, SET TRANSACTION ISOLATION LEVEL,
10615 SELECT ... FOR UPDATE, and an improved LOCK TABLE command.
10626 We continue to speed up <productname>PostgreSQL</productname>,
10627 thanks to the variety of talents within our team. We have
10628 sped up memory allocation, optimization, table joins, and row
10640 We continue to expand our port list, this time including
10641 <systemitem class="osname">Windows NT</>/<systemitem>ix86</> and <systemitem class="osname">NetBSD</>/<systemitem>arm32</>.
10652 Most interfaces have new versions, and existing functionality
10664 New and updated material is present throughout the
10665 documentation. New <acronym>FAQ</acronym>s have been
10666 contributed for <systemitem class="osname">SGI</> and <systemitem class="osname">AIX</> platforms.
10667 The <citetitle>Tutorial</citetitle> has introductory information
10668 on <acronym>SQL</acronym> from Stefan Simkovics.
10669 For the <citetitle>User's Guide</citetitle>, there are
10670 reference pages covering the postmaster and more utility
10671 programs, and a new appendix
10672 contains details on date/time behavior.
10673 The <citetitle>Administrator's Guide</citetitle> has a new
10674 chapter on troubleshooting from Tom Lane.
10675 And the <citetitle>Programmer's Guide</citetitle> has a
10676 description of query processing, also from Stefan, and details
10677 on obtaining the <productname>PostgreSQL</productname> source
10678 tree via anonymous <productname>CVS</productname> and
10679 <productname>CVSup</productname>.
10687 <title>Migration to version 6.5</title>
10690 A dump/restore using <application>pg_dump</application>
10691 is required for those wishing to migrate data from any
10692 previous release of <productname>PostgreSQL</productname>.
10693 <application>pg_upgrade</application> can <emphasis>not</emphasis>
10694 be used to upgrade to this release because the on-disk structure
10695 of the tables has changed compared to previous releases.
10699 The new Multiversion Concurrency Control (MVCC) features can
10700 give somewhat different behaviors in multiuser
10701 environments. <emphasis>Read and understand the following section
10702 to ensure that your existing applications will give you the
10703 behavior you need.</emphasis>
10707 <title>Multiversion Concurrency Control</title>
10710 Because readers in 6.5 don't lock data, regardless of transaction
10711 isolation level, data read by one transaction can be overwritten by
10712 another. In other words, if a row is returned by
10713 <command>SELECT</command> it doesn't mean that this row really exists
10714 at the time it is returned (i.e. sometime after the statement or
10715 transaction began) nor that the row is protected from being deleted or
10716 updated by concurrent transactions before the current transaction does
10717 a commit or rollback.
10721 To ensure the actual existence of a row and protect it against
10722 concurrent updates one must use <command>SELECT FOR UPDATE</command> or
10723 an appropriate <command>LOCK TABLE</command> statement. This should be
10724 taken into account when porting applications from previous releases of
10725 <productname>PostgreSQL</productname> and other environments.
10729 Keep the above in mind if you are using
10730 <filename>contrib/refint.*</filename> triggers for
10731 referential integrity. Additional techniques are required now. One way is
10732 to use <command>LOCK parent_table IN SHARE ROW EXCLUSIVE MODE</command>
10733 command if a transaction is going to update/delete a primary key and
10734 use <command>LOCK parent_table IN SHARE MODE</command> command if a
10735 transaction is going to update/insert a foreign key.
10739 Note that if you run a transaction in SERIALIZABLE mode then you must
10740 execute the <command>LOCK</command> commands above before execution of any
10741 <acronym>DML</acronym> statement
10742 (<command>SELECT/INSERT/DELETE/UPDATE/FETCH/COPY_TO</command>) in the
10749 These inconveniences will disappear in the future
10750 when the ability to read dirty
10751 (uncommitted) data (regardless of isolation level) and true referential
10752 integrity will be implemented.
10758 <title>Changes</title>
10764 Fix text<->float8 and text<->float4 conversion functions(Thomas)
10765 Fix for creating tables with mixed-case constraints(Billy)
10766 Change exp()/pow() behavior to generate error on underflow/overflow(Jan)
10767 Fix bug in pg_dump -z
10768 Memory overrun cleanups(Tatsuo)
10769 Fix for lo_import crash(Tatsuo)
10770 Adjust handling of data type names to suppress double quotes(Thomas)
10771 Use type coercion for matching columns and DEFAULT(Thomas)
10772 Fix deadlock so it only checks once after one second of sleep(Bruce)
10773 Fixes for aggregates and PL/pgsql(Hiroshi)
10774 Fix for subquery crash(Vadim)
10775 Fix for libpq function PQfnumber and case-insensitive names(Bahman Rafatjoo)
10776 Fix for large object write-in-middle, no extra block, memory consumption(Tatsuo)
10777 Fix for pg_dump -d or -D and quote special characters in INSERT
10778 Repair serious problems with dynahash(Tom)
10779 Fix INET/CIDR portability problems
10780 Fix problem with selectivity error in ALTER TABLE ADD COLUMN(Bruce)
10781 Fix executor so mergejoin of different column types works(Tom)
10782 Fix for Alpha OR selectivity bug
10783 Fix OR index selectivity problem(Bruce)
10784 Fix so \d shows proper length for char()/varchar()(Ryan)
10785 Fix tutorial code(Clark)
10786 Improve destroyuser checking(Oliver)
10787 Fix for Kerberos(Rodney McDuff)
10788 Fix for dropping database while dirty buffers(Bruce)
10789 Fix so sequence nextval() can be case-sensitive(Bruce)
10791 Drop buffers before destroying database files(Bruce)
10792 Fix case where executor evaluates functions twice(Tatsuo)
10793 Allow sequence nextval actions to be case-sensitive(Bruce)
10794 Fix optimizer indexing not working for negative numbers(Bruce)
10795 Fix for memory leak in executor with fjIsNull
10796 Fix for aggregate memory leaks(Erik Riedel)
10797 Allow user name containing a dash to grant privileges
10798 Cleanup of NULL in inet types
10799 Clean up system table bugs(Tom)
10800 Fix problems of PAGER and \? command(Masaaki Sakaida)
10801 Reduce default multisegment file size limit to 1GB(Peter)
10802 Fix for dumping of CREATE OPERATOR(Tom)
10803 Fix for backward scanning of cursors(Hiroshi Inoue)
10804 Fix for COPY FROM STDIN when using \i(Tom)
10805 Fix for subselect is compared inside an expression(Jan)
10806 Fix handling of error reporting while returning rows(Tom)
10807 Fix problems with reference to array types(Tom,Jan)
10808 Prevent UPDATE SET oid(Jan)
10809 Fix pg_dump so -t option can handle case-sensitive tablenames
10810 Fixes for GROUP BY in special cases(Tom, Jan)
10811 Fix for memory leak in failed queries(Tom)
10812 DEFAULT now supports mixed-case identifiers(Tom)
10813 Fix for multisegment uses of DROP/RENAME table, indexes(Ole Gjerde)
10814 Disable use of pg_dump with both -o and -d options(Bruce)
10815 Allow pg_dump to properly dump group privileges(Bruce)
10816 Fix GROUP BY in INSERT INTO table SELECT * FROM table2(Jan)
10817 Fix for computations in views(Jan)
10818 Fix for aggregates on array indexes(Tom)
10819 Fix for DEFAULT handles single quotes in value requiring too many quotes
10820 Fix security problem with non-super users importing/exporting large objects(Tom)
10821 Rollback of transaction that creates table cleaned up properly(Tom)
10822 Fix to allow long table and column names to generate proper serial names(Tom)
10826 Add "vacuumdb" utility
10827 Speed up libpq by allocating memory better(Tom)
10828 EXPLAIN all indexes used(Tom)
10829 Implement CASE, COALESCE, NULLIF expression(Thomas)
10830 New pg_dump table output format(Constantin)
10831 Add string min()/max() functions(Thomas)
10832 Extend new type coercion techniques to aggregates(Thomas)
10833 New moddatetime contrib(Terry)
10834 Update to pgaccess 0.96(Constantin)
10835 Add routines for single-byte "char" type(Thomas)
10836 Improved substr() function(Thomas)
10837 Improved multibyte handling(Tatsuo)
10838 Multiversion concurrency control/MVCC(Vadim)
10839 New Serialized mode(Vadim)
10840 Fix for tables over 2gigs(Peter)
10841 New SET TRANSACTION ISOLATION LEVEL(Vadim)
10842 New LOCK TABLE IN ... MODE(Vadim)
10843 Update ODBC driver(Byron)
10844 New NUMERIC data type(Jan)
10845 New SELECT FOR UPDATE(Vadim)
10846 Handle "NaN" and "Infinity" for input values(Jan)
10847 Improved date/year handling(Thomas)
10848 Improved handling of backend connections(Magnus)
10849 New options ELOG_TIMESTAMPS and USE_SYSLOG options for log files(Massimo)
10850 New TCL_ARRAYS option(Massimo)
10851 New INTERSECT and EXCEPT(Stefan)
10852 New pg_index.indisprimary for primary key tracking(D'Arcy)
10853 New pg_dump option to allow dropping of tables before creation(Brook)
10854 Speedup of row output routines(Tom)
10855 New READ COMMITTED isolation level(Vadim)
10856 New TEMP tables/indexes(Bruce)
10857 Prevent sorting if result is already sorted(Jan)
10858 New memory allocation optimization(Jan)
10859 Allow psql to do \p\g(Bruce)
10860 Allow multiple rule actions(Jan)
10861 Added LIMIT/OFFSET functionality(Jan)
10862 Improve optimizer when joining a large number of tables(Bruce)
10863 New intro to SQL from S. Simkovics' Master's Thesis (Stefan, Thomas)
10864 New intro to backend processing from S. Simkovics' Master's Thesis (Stefan)
10865 Improved int8 support(Ryan Bradetich, Thomas, Tom)
10866 New routines to convert between int8 and text/varchar types(Thomas)
10867 New bushy plans, where meta-tables are joined(Bruce)
10868 Enable right-hand queries by default(Bruce)
10869 Allow reliable maximum number of backends to be set at configure time
10870 (--with-maxbackends and postmaster switch (-N backends))(Tom)
10871 GEQO default now 10 tables because of optimizer speedups(Tom)
10872 Allow NULL=Var for MS-SQL portability(Michael, Bruce)
10873 Modify contrib check_primary_key() so either "automatic" or "dependent"(Anand)
10874 Allow psql \d on a view show query(Ryan)
10875 Speedup for LIKE(Bruce)
10876 Ecpg fixes/features, see src/interfaces/ecpg/ChangeLog file(Michael)
10877 JDBC fixes/features, see src/interfaces/jdbc/CHANGELOG(Peter)
10878 Make % operator have precedence like /(Bruce)
10879 Add new postgres -O option to allow system table structure changes(Bruce)
10880 Update contrib/pginterface/findoidjoins script(Tom)
10881 Major speedup in vacuum of deleted rows with indexes(Vadim)
10882 Allow non-SQL functions to run different versions based on arguments(Tom)
10883 Add -E option that shows actual queries sent by \dt and friends(Masaaki Sakaida)
10884 Add version number in start-up banners for psql(Masaaki Sakaida)
10885 New contrib/vacuumlo removes large objects not referenced(Peter)
10886 New initialization for table sizes so non-vacuumed tables perform better(Tom)
10887 Improve error messages when a connection is rejected(Tom)
10888 Support for arrays of char() and varchar() fields(Massimo)
10889 Overhaul of hash code to increase reliability and performance(Tom)
10890 Update to PyGreSQL 2.4(D'Arcy)
10891 Changed debug options so -d4 and -d5 produce different node displays(Jan)
10892 New pg_options: pretty_plan, pretty_parse, pretty_rewritten(Jan)
10893 Better optimization statistics for system table access(Tom)
10894 Better handling of non-default block sizes(Massimo)
10895 Improve GEQO optimizer memory consumption(Tom)
10896 UNION now suppports ORDER BY of columns not in target list(Jan)
10897 Major libpq++ improvements(Vince Vielhaber)
10898 pg_dump now uses -z(ACL's) as default(Bruce)
10899 backend cache, memory speedups(Tom)
10900 have pg_dump do everything in one snapshot transaction(Vadim)
10901 fix for large object memory leakage, fix for pg_dumping(Tom)
10902 INET type now respects netmask for comparisons
10903 Make VACUUM ANALYZE only use a readlock(Vadim)
10904 Allow VIEWs on UNIONS(Jan)
10905 pg_dump now can generate consistent snapshots on active databases(Vadim)
10907 Source Tree Changes
10908 -------------------
10909 Improve port matching(Tom)
10910 Portability fixes for SunOS
10911 Add Windows NT backend port and enable dynamic loading(Magnus and Daniel Horak)
10912 New port to Cobalt Qube(Mips) running Linux(Tatsuo)
10913 Port to NetBSD/m68k(Mr. Mutsuki Nakajima)
10914 Port to NetBSD/sun3(Mr. Mutsuki Nakajima)
10915 Port to NetBSD/macppc(Toshimi Aoki)
10916 Fix for tcl/tk configuration(Vince)
10917 Removed CURRENT key word for rule queries(Jan)
10918 NT dynamic loading now works(Daniel Horak)
10919 Add ARM32 support(Andrew McMurry)
10920 Better support for HP-UX 11 and UnixWare
10921 Improve file handling to be more uniform, prevent file descriptor leak(Tom)
10922 New install commands for plpgsql(Jan)
10929 <sect1 id="release-6-4-2">
10930 <title>Release 6.4.2</title>
10933 <title>Release date</title>
10934 <simpara>1998-12-20</simpara>
10938 The 6.4.1 release was improperly packaged. This also has one additional
10944 <title>Migration to version 6.4.2</title>
10947 A dump/restore is <emphasis>not</emphasis> required for those running
10952 <title>Changes</title>
10956 Fix for datetime constant problem on some platforms(Thomas)
10964 <sect1 id="release-6-4-1">
10965 <title>Release 6.4.1</title>
10968 <title>Release date</title>
10969 <simpara>1998-12-18</simpara>
10973 This is basically a cleanup release for 6.4. We have fixed a variety of
10974 problems reported by 6.4 users.
10979 <title>Migration to version 6.4.1</title>
10982 A dump/restore is <emphasis>not</emphasis> required for those running
10987 <title>Changes</title>
10991 Add pg_dump -N flag to force double quotes around identifiers. This is
10992 the default(Thomas)
10993 Fix for NOT in where clause causing crash(Bruce)
10994 EXPLAIN VERBOSE coredump fix(Vadim)
10995 Fix shared-library problems on Linux
10996 Fix test for table existence to allow mixed-case and whitespace in
10997 the table name(Thomas)
10998 Fix a couple of pg_dump bugs
10999 Configure matches template/.similar entries better(Tom)
11000 Change builtin function names from SPI_* to spi_*
11001 OR WHERE clause fix(Vadim)
11002 Fixes for mixed-case table names(Billy)
11003 contrib/linux/postgres.init.csh/sh fix(Thomas)
11004 libpq memory overrun fix
11006 Change exp() behavior to generate error on underflow(Thomas)
11007 pg_dump fixes for memory leak, inheritance constraints, layout change
11008 update pgaccess to 0.93
11009 Fix prototype for 64-bit platforms
11010 Multibyte fixes(Tatsuo)
11012 Fix memory overruns(Tatsuo)
11013 Fix for lo_import() crash(Bruce)
11014 Better search for install program(Tom)
11015 Timezone fixes(Tom)
11017 Use implicit type coercion for matching DEFAULT values(Thomas)
11018 Add routines to help with single-byte (internal) character type(Thomas)
11019 Compilation of libpq for Windows fixes(Magnus)
11020 Upgrade to PyGreSQL 2.2(D'Arcy)
11028 <sect1 id="release-6-4">
11029 <title>Release 6.4</title>
11032 <title>Release date</title>
11033 <simpara>1998-10-30</simpara>
11037 There are <emphasis>many</emphasis> new features and improvements in this release.
11038 Thanks to our developers and maintainers, nearly every aspect of the system
11039 has received some attention since the previous release.
11040 Here is a brief, incomplete summary:
11045 Views and rules are now functional thanks to extensive new code in the
11046 rewrite rules system from Jan Wieck. He also wrote a chapter on it
11047 for the <citetitle>Programmer's Guide</citetitle>.
11052 Jan also contributed a second procedural language, <application>PL/pgSQL</application>, to go with the
11053 original <application>PL/pgTCL</application> procedural language he contributed last release.
11059 We have optional multiple-byte character set support from Tatsuo Ishii
11060 to complement our existing locale support.
11066 Client/server communications has been cleaned up, with better support for
11067 asynchronous messages and interrupts thanks to Tom Lane.
11073 The parser will now perform automatic type coercion to match arguments
11074 to available operators and functions, and to match columns and expressions
11075 with target columns. This uses a generic mechanism which supports
11076 the type extensibility features of <productname>PostgreSQL</productname>.
11077 There is a new chapter in the <citetitle>User's Guide</citetitle>
11078 which covers this topic.
11084 Three new data types have been added.
11085 Two types, <type>inet</type> and <type>cidr</type>, support various forms
11086 of IP network, subnet, and machine addressing. There is now an 8-byte integer
11087 type available on some platforms. See the chapter on data types
11088 in the <citetitle>User's Guide</citetitle> for details.
11089 A fourth type, <type>serial</type>, is now supported by the parser as an
11090 amalgam of the <type>int4</type> type, a sequence, and a unique index.
11096 Several more <acronym>SQL92</acronym>-compatible syntax features have been
11097 added, including <command>INSERT DEFAULT VALUES</command>
11103 The automatic configuration and installation system has received some
11104 attention, and should be more robust for more platforms than it has ever
11113 <title>Migration to version 6.4</title>
11116 A dump/restore using <application>pg_dump</application>
11117 or <application>pg_dumpall</application>
11118 is required for those wishing to migrate data from any
11119 previous release of <productname>PostgreSQL</productname>.
11124 <title>Changes</title>
11130 Fix for a tiny memory leak in PQsetdb/PQfinish(Bryan)
11131 Remove char2-16 data types, use char/varchar(Darren)
11132 Pqfn not handles a NOTICE message(Anders)
11133 Reduced busywaiting overhead for spinlocks with many backends (dg)
11134 Stuck spinlock detection (dg)
11135 Fix up "ISO-style" timespan decoding and encoding(Thomas)
11136 Fix problem with table drop after rollback of transaction(Vadim)
11137 Change error message and remove non-functional update message(Vadim)
11138 Fix for COPY array checking
11139 Fix for SELECT 1 UNION SELECT NULL
11140 Fix for buffer leaks in large object calls(Pascal)
11141 Change owner from oid to int4 type(Bruce)
11142 Fix a bug in the oracle compatibility functions btrim() ltrim() and rtrim()
11143 Fix for shared invalidation cache overflow(Massimo)
11144 Prevent file descriptor leaks in failed COPY's(Bruce)
11145 Fix memory leak in libpgtcl's pg_select(Constantin)
11146 Fix problems with username/passwords over 8 characters(Tom)
11147 Fix problems with handling of asynchronous NOTIFY in backend(Tom)
11148 Fix of many bad system table entries(Tom)
11152 Upgrade ecpg and ecpglib,see src/interfaces/ecpc/ChangeLog(Michael)
11153 Show the index used in an EXPLAIN(Zeugswetter)
11154 EXPLAIN invokes rule system and shows plan(s) for rewritten queries(Jan)
11155 Multibyte awareness of many data types and functions, via configure(Tatsuo)
11156 New configure --with-mb option(Tatsuo)
11157 New initdb --pgencoding option(Tatsuo)
11158 New createdb -E multibyte option(Tatsuo)
11159 Select version(); now returns PostgreSQL version(Jeroen)
11160 libpq now allows asynchronous clients(Tom)
11161 Allow cancel from client of backend query(Tom)
11162 psql now cancels query with Control-C(Tom)
11163 libpq users need not issue dummy queries to get NOTIFY messages(Tom)
11164 NOTIFY now sends sender's PID, so you can tell whether it was your own(Tom)
11165 PGresult struct now includes associated error message, if any(Tom)
11166 Define "tz_hour" and "tz_minute" arguments to date_part()(Thomas)
11167 Add routines to convert between varchar and bpchar(Thomas)
11168 Add routines to allow sizing of varchar and bpchar into target columns(Thomas)
11169 Add bit flags to support timezonehour and minute in data retrieval(Thomas)
11170 Allow more variations on valid floating point numbers (e.g. ".1", "1e6")(Thomas)
11171 Fixes for unary minus parsing with leading spaces(Thomas)
11172 Implement TIMEZONE_HOUR, TIMEZONE_MINUTE per SQL92 specs(Thomas)
11173 Check for and properly ignore FOREIGN KEY column constraints(Thomas)
11174 Define USER as synonym for CURRENT_USER per SQL92 specs(Thomas)
11175 Enable HAVING clause but no fixes elsewhere yet.
11176 Make "char" type a synonym for "char(1)" (actually implemented as bpchar)(Thomas)
11177 Save string type if specified for DEFAULT clause handling(Thomas)
11178 Coerce operations involving different data types(Thomas)
11179 Allow some index use for columns of different types(Thomas)
11180 Add capabilities for automatic type conversion(Thomas)
11181 Cleanups for large objects, so file is truncated on open(Peter)
11182 Readline cleanups(Tom)
11183 Allow psql \f \ to make spaces as delimiter(Bruce)
11184 Pass pg_attribute.atttypmod to the frontend for column field lengths(Tom,Bruce)
11185 Msql compatibility library in /contrib(Aldrin)
11186 Remove the requirement that ORDER/GROUP BY clause identifiers be
11187 included in the target list(David)
11188 Convert columns to match columns in UNION clauses(Thomas)
11189 Remove fork()/exec() and only do fork()(Bruce)
11190 Jdbc cleanups(Peter)
11191 Show backend status on ps command line(only works on some platforms)(Bruce)
11192 Pg_hba.conf now has a sameuser option in the database field
11193 Make lo_unlink take oid param, not int4
11194 New DISABLE_COMPLEX_MACRO for compilers that can't handle our macros(Bruce)
11195 Libpgtcl now handles NOTIFY as a Tcl event, need not send dummy queries(Tom)
11196 libpgtcl cleanups(Tom)
11197 Add -error option to libpgtcl's pg_result command(Tom)
11198 New locale patch, see docs/README/locale(Oleg)
11199 Fix for pg_dump so CONSTRAINT and CHECK syntax is correct(ccb)
11200 New contrib/lo code for large object orphan removal(Peter)
11201 New psql command "SET CLIENT_ENCODING TO 'encoding'" for multibytes
11202 feature, see /doc/README.mb(Tatsuo)
11203 contrib/noupdate code to revoke update permission on a column
11204 libpq can now be compiled on Windows(Magnus)
11205 Add PQsetdbLogin() in libpq
11206 New 8-byte integer type, checked by configure for OS support(Thomas)
11207 Better support for quoted table/column names(Thomas)
11208 Surround table and column names with double-quotes in pg_dump(Thomas)
11209 PQreset() now works with passwords(Tom)
11210 Handle case of GROUP BY target list column number out of range(David)
11211 Allow UNION in subselects
11212 Add auto-size to screen to \d? commands(Bruce)
11213 Use UNION to show all \d? results in one query(Bruce)
11214 Add \d? field search feature(Bruce)
11215 Pg_dump issues fewer \connect requests(Tom)
11216 Make pg_dump -z flag work better, document it in manual page(Tom)
11217 Add HAVING clause with full support for subselects and unions(Stephan)
11218 Full text indexing routines in contrib/fulltextindex(Maarten)
11219 Transaction ids now stored in shared memory(Vadim)
11220 New PGCLIENTENCODING when issuing COPY command(Tatsuo)
11221 Support for SQL92 syntax "SET NAMES"(Tatsuo)
11222 Support for LATIN2-5(Tatsuo)
11223 Add UNICODE regression test case(Tatsuo)
11224 Lock manager cleanup, new locking modes for LLL(Vadim)
11225 Allow index use with OR clauses(Bruce)
11226 Allows "SELECT NULL ORDER BY 1;"
11227 Explain VERBOSE prints the plan, and now pretty-prints the plan to
11228 the postmaster log file(Bruce)
11229 Add indexes display to \d command(Bruce)
11230 Allow GROUP BY on functions(David)
11231 New pg_class.relkind for large objects(Bruce)
11232 New way to send libpq NOTICE messages to a different location(Tom)
11233 New \w write command to psql(Bruce)
11234 New /contrib/findoidjoins scans oid columns to find join relationships(Bruce)
11235 Allow binary-compatible indexes to be considered when checking for valid
11236 Indexes for restriction clauses containing a constant(Thomas)
11237 New ISBN/ISSN code in /contrib/isbn_issn
11238 Allow NOT LIKE, IN, NOT IN, BETWEEN, and NOT BETWEEN constraint(Thomas)
11239 New rewrite system fixes many problems with rules and views(Jan)
11240 * Rules on relations work
11241 * Event qualifications on insert/update/delete work
11242 * New OLD variable to reference CURRENT, CURRENT will be remove in future
11243 * Update rules can reference NEW and OLD in rule qualifications/actions
11244 * Insert/update/delete rules on views work
11245 * Multiple rule actions are now supported, surrounded by parentheses
11246 * Regular users can create views/rules on tables they have RULE permits
11247 * Rules and views inherit the privileges of the creator
11248 * No rules at the column level
11249 * No UPDATE NEW/OLD rules
11250 * New pg_tables, pg_indexes, pg_rules and pg_views system views
11251 * Only a single action on SELECT rules
11252 * Total rewrite overhaul, perhaps for 6.5
11253 * handle subselects
11254 * handle aggregates on views
11255 * handle insert into select from view works
11256 System indexes are now multikey(Bruce)
11257 Oidint2, oidint4, and oidname types are removed(Bruce)
11258 Use system cache for more system table lookups(Bruce)
11259 New backend programming language PL/pgSQL in backend/pl(Jan)
11260 New SERIAL data type, auto-creates sequence/index(Thomas)
11261 Enable assert checking without a recompile(Massimo)
11262 User lock enhancements(Massimo)
11263 New setval() command to set sequence value(Massimo)
11264 Auto-remove unix socket file on start-up if no postmaster running(Massimo)
11265 Conditional trace package(Massimo)
11266 New UNLISTEN command(Massimo)
11267 psql and libpq now compile under Windows using win32.mak(Magnus)
11268 Lo_read no longer stores trailing NULL(Bruce)
11269 Identifiers are now truncated to 31 characters internally(Bruce)
11270 Createuser options now availble on the command line
11271 Code for 64-bit integer supported added, configure tested, int8 type(Thomas)
11272 Prevent file descriptor leaf from failed COPY(Bruce)
11273 New pg_upgrade command(Bruce)
11274 Updated /contrib directories(Massimo)
11275 New CREATE TABLE DEFAULT VALUES statement available(Thomas)
11276 New INSERT INTO TABLE DEFAULT VALUES statement available(Thomas)
11277 New DECLARE and FETCH feature(Thomas)
11278 libpq's internal structures now not exported(Tom)
11279 Allow up to 8 key indexes(Bruce)
11280 Remove ARCHIVE key word, that is no longer used(Thomas)
11281 pg_dump -n flag to supress quotes around indentifiers
11282 disable system columns for views(Jan)
11283 new INET and CIDR types for network addresses(TomH, Paul)
11284 no more double quotes in psql output
11285 pg_dump now dumps views(Terry)
11286 new SET QUERY_LIMIT(Tatsuo,Jan)
11288 Source Tree Changes
11289 -------------------
11290 /contrib cleanup(Jun)
11291 Inline some small functions called for every row(Bruce)
11293 HP-UX cleanups(Tom)
11294 Multibyte regression tests(Soonmyung.)
11295 Remove --disabled options from configure
11296 Define PGDOC to use POSTGRESDIR by default
11297 Make regression optional
11298 Remove extra braces code to pgindent(Bruce)
11299 Add bsdi shared library support(Bruce)
11300 New --without-CXX support configure option(Brook)
11302 Update backend flowchart in tools/backend(Bruce)
11303 Change atttypmod from int16 to int32(Bruce, Tom)
11304 Getrusage() fix for platforms that do not have it(Tom)
11305 Add PQconnectdb, PGUSER, PGPASSWORD to libpq man page
11306 NS32K platform fixes(Phil Nelson, John Buller)
11307 SCO 7/UnixWare 2.x fixes(Billy,others)
11308 Sparc/Solaris 2.5 fixes(Ryan)
11309 Pgbuiltin.3 is obsolete, move to doc files(Thomas)
11310 Even more documention(Thomas)
11311 Nextstep support(Jacek)
11313 pginterface manual page(Bruce)
11314 shared libraries all have version numbers
11315 merged all OS-specific shared library defines into one file
11316 smarter TCL/TK configuration checking(Billy)
11317 smarter perl configuration(Brook)
11318 configure uses supplied install-sh if no install script found(Tom)
11319 new Makefile.shlib for shared library configuration(Tom)
11325 <sect1 id="release-6-3-2">
11326 <title>Release 6.3.2</title>
11329 <title>Release date</title>
11330 <simpara>1998-04-07</simpara>
11334 This is a bug-fix release for 6.3.x.
11335 Refer to the release notes for version 6.3 for a more complete summary of new features.
11343 Repairs automatic configuration support for some platforms, including Linux,
11344 from breakage inadvertently introduced in version 6.3.1.
11350 Correctly handles function calls on the left side of BETWEEN and LIKE clauses.
11357 A dump/restore is NOT required for those running 6.3 or 6.3.1. A
11358 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
11359 This last step should be performed while the postmaster is not running.
11360 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
11363 For upgrades from pre-6.3 installations,
11364 refer to the installation and migration instructions for version 6.3.
11368 <title>Changes</title>
11372 Configure detection improvements for tcl/tk(Brook Milligan, Alvin)
11373 Manual page improvements(Bruce)
11374 BETWEEN and LIKE fix(Thomas)
11375 fix for psql \connect used by pg_dump(Oliver Elphick)
11377 pgaccess, version 0.86
11378 qsort removed, now uses libc version, cleanups(Jeroen)
11379 fix for buffer over-runs detected(Maurice Gittens)
11380 fix for buffer overrun in libpgtcl(Randy Kunkee)
11381 fix for UNION with DISTINCT or ORDER BY(Bruce)
11382 gettimeofday configure check(Doug Winterburn)
11383 Fix "indexes not used" bug(Vadim)
11384 docs additions(Thomas)
11385 Fix for backend memory leak(Bruce)
11386 libreadline cleanup(Erwan MAS)
11387 Remove DISTDIR(Bruce)
11388 Makefile dependency cleanup(Jeroen van Vianen)
11389 ASSERT fixes(Bruce)
11395 <sect1 id="release-6-3-1">
11396 <title>Release 6.3.1</title>
11399 <title>Release date</title>
11400 <simpara>1998-03-23</simpara>
11409 Additional support for multibyte character sets.
11415 Repair byte ordering for mixed-endian clients and servers.
11421 Minor updates to allowed SQL syntax.
11427 Improvements to the configuration autodetection for installation.
11434 A dump/restore is NOT required for those running 6.3. A
11435 <literal>make distclean</>, <literal>make</>, and <literal>make install</> is all that is required.
11436 This last step should be performed while the postmaster is not running.
11437 You should re-link any custom applications that use <productname>PostgreSQL</productname> libraries.
11440 For upgrades from pre-6.3 installations,
11441 refer to the installation and migration instructions for version 6.3.
11445 <title>Changes</title>
11449 ecpg cleanup/fixes, now version 1.1(Michael Meskes)
11450 pg_user cleanup(Bruce)
11451 large object fix for pg_dump and tclsh (alvin)
11452 LIKE fix for multiple adjacent underscores
11453 fix for redefining builtin functions(Thomas)
11455 upgrade to pg_access 0.83
11456 updated CLUSTER manual page
11457 multibyte character set support, see doc/README.mb(Tatsuo)
11458 configure --with-pgport fix
11460 big-endian fix for backend communications(Kataoka)
11461 SUBSTR() and substring() fix(Jan)
11462 several jdbc fixes(Peter)
11463 libpgtcl improvements, see libptcl/README(Randy Kunkee)
11464 Fix for "Datasize = 0" error(Vadim)
11465 Prevent \do from wrapping(Bruce)
11466 Remove duplicate Russian character set entries
11468 Allow optional TABLE key word in LOCK and SELECT INTO(Thomas)
11469 CREATE SEQUENCE options to allow a negative integer(Thomas)
11470 Add "PASSWORD" as an allowed column identifier(Thomas)
11471 Add checks for UNION target fields(Bruce)
11472 Fix Alpha port(Dwayne Bailey)
11473 Fix for text arrays containing quotes(Doug Gibson)
11474 Solaris compile fix(Albert Chin-A-Young)
11475 Better identify tcl and tk libs and includes(Bruce)
11481 <sect1 id="release-6-3">
11482 <title>Release 6.3</title>
11485 <title>Release date</title>
11486 <simpara>1998-03-01</simpara>
11490 There are <emphasis>many</emphasis> new features and improvements in this release.
11491 Here is a brief, incomplete summary:
11496 Many new SQL features, including
11497 full <acronym>SQL92</acronym> subselect capability
11498 (everything is here but target-list subselects).
11504 Support for client-side environment variables to specify time zone and date style.
11510 Socket interface for client/server connection. This is the default now
11511 so you may need to start <application>postmaster</application> with the
11512 <option>-i</option> flag.
11518 Better password authorization mechanisms. Default table privileges have changed.
11524 Old-style <firstterm>time travel</firstterm>
11525 has been removed. Performance has been improved.
11533 Bruce Momjian wrote the following notes to introduce the new release.
11538 There are some general 6.3 issues that I want to mention. These are
11539 only the big items that cannot be described in one sentence. A review
11540 of the detailed changes list is still needed.
11543 First, we now have subselects. Now that we have them, I would like to
11544 mention that without subselects, SQL is a very limited language.
11545 Subselects are a major feature, and you should review your code for
11546 places where subselects provide a better solution for your queries. I
11547 think you will find that there are more uses for subselects than you may
11548 think. Vadim has put us on the big SQL map with subselects, and fully
11549 functional ones too. The only thing you can't do with subselects is to
11550 use them in the target list.
11553 Second, 6.3 uses Unix domain sockets rather than TCP/IP by default. To
11554 enable connections from other machines, you have to use the new
11555 postmaster -i option, and of course edit <filename>pg_hba.conf</filename>. Also, for this
11556 reason, the format of <filename>pg_hba.conf</filename> has changed.
11559 Third, <type>char()</type> fields will now allow faster access than <type>varchar()</type> or
11560 <type>text</type>. Specifically, the <type>text</> and <type>varchar()</type> have a penalty for access to
11561 any columns after the first column of this type. <type>char()</type> used to also
11562 have this access penalty, but it no longer does. This may suggest that
11563 you redesign some of your tables, especially if you have short character
11564 columns that you have defined as <type>varchar()</type> or <type>text</type>. This and other
11565 changes make 6.3 even faster than earlier releases.
11568 We now have passwords definable independent of any Unix file. There are
11569 new SQL USER commands.
11570 See the <citetitle>Administrator's Guide</citetitle> for more
11571 information. There is a new table, pg_shadow, which is used to store
11572 user information and user passwords, and it by default only SELECT-able
11573 by the <systemitem>postgres</systemitem> super-user. pg_user is now a view of pg_shadow, and is
11574 SELECT-able by PUBLIC. You should keep using pg_user in your
11575 application without changes.
11578 User-created tables now no longer have SELECT privilege to PUBLIC by
11579 default. This was done because the ANSI standard requires it. You can
11580 of course GRANT any privileges you want after the table is created.
11581 System tables continue to be SELECT-able by PUBLIC.
11584 We also have real deadlock detection code. No more sixty-second
11585 timeouts. And the new locking code implements a <acronym>FIFO</acronym> better, so there
11586 should be less resource starvation during heavy use.
11589 Many complaints have been made about inadequate documentation in previous
11590 releases. Thomas has put much effort into many new manuals for this
11591 release. Check out the doc/ directory.
11594 For performance reasons, time travel is gone, but can be implemented
11595 using triggers (see <filename>pgsql/contrib/spi/README</filename>). Please check out the new
11596 \d command for types, operators, etc. Also, views have their own
11597 privileges now, not based on the underlying tables, so privileges on
11598 them have to be set separately. Check <filename>/pgsql/interfaces</filename> for some new
11599 ways to talk to <productname>PostgreSQL</productname>.
11602 This is the first release that really required an explanation for
11603 existing users. In many ways, this was necessary because the new
11604 release removes many limitations, and the work-arounds people were using
11605 are no longer needed.
11609 <title>Migration to version 6.3</title>
11612 A dump/restore using <application>pg_dump</application>
11613 or <application>pg_dumpall</application>
11614 is required for those wishing to migrate data from any
11615 previous release of <productname>PostgreSQL</productname>.
11620 <title>Changes</title>
11626 Fix binary cursors broken by MOVE implementation(Vadim)
11627 Fix for tcl library crash(Jan)
11628 Fix for array handling, from Gerhard Hintermayer
11629 Fix acl error, and remove duplicate pqtrace(Bruce)
11630 Fix psql \e for empty file(Bruce)
11631 Fix for textcat on varchar() fields(Bruce)
11632 Fix for DBT Sendproc (Zeugswetter Andres)
11633 Fix vacuum analyze syntax problem(Bruce)
11634 Fix for international identifiers(Tatsuo)
11635 Fix aggregates on inherited tables(Bruce)
11636 Fix substr() for out-of-bounds data
11637 Fix for select 1=1 or 2=2, select 1=1 and 2=2, and select sum(2+2)(Bruce)
11638 Fix notty output to show status result. -q option still turns it off(Bruce)
11639 Fix for count(*), aggs with views and multiple tables and sum(3)(Bruce)
11641 Fix for PQtrace start/stop several times(Bruce)
11642 Fix a variety of locking problems like newer lock waiters getting
11643 lock before older waiters, and having readlock people not share
11644 locks if a writer is waiting for a lock, and waiting writers not
11645 getting priority over waiting readers(Bruce)
11646 Fix crashes in psql when executing queries from external files(James)
11647 Fix problem with multiple order by columns, with the first one having
11648 NULL values(Jeroen)
11649 Use correct hash table support functions for float8 and int4(Thomas)
11650 Re-enable JOIN= option in CREATE OPERATOR statement (Thomas)
11651 Change precedence for boolean operators to match expected behavior(Thomas)
11652 Generate elog(ERROR) on over-large integer(Bruce)
11653 Allow multiple-argument functions in constraint clauses(Thomas)
11654 Check boolean input literals for 'true','false','yes','no','1','0'
11655 and throw elog(ERROR) if unrecognized(Thomas)
11656 Major large objects fix
11657 Fix for GROUP BY showing duplicates(Vadim)
11658 Fix for index scans in MergeJion(Vadim)
11662 Subselects with EXISTS, IN, ALL, ANY key words (Vadim, Bruce, Thomas)
11663 New User Manual(Thomas, others)
11664 Speedup by inlining some frequently-called functions
11665 Real deadlock detection, no more timeouts(Bruce)
11666 Add SQL92 "constants" CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP,
11667 CURRENT_USER(Thomas)
11668 Modify constraint syntax to be SQL92-compliant(Thomas)
11669 Implement SQL92 PRIMARY KEY and UNIQUE clauses using indexes(Thomas)
11670 Recognize SQL92 syntax for FOREIGN KEY. Throw elog notice(Thomas)
11671 Allow NOT NULL UNIQUE constraint clause (each allowed separately before)(Thomas)
11672 Allow PostgreSQL-style casting ("::") of non-constants(Thomas)
11673 Add support for SQL3 TRUE and FALSE boolean constants(Thomas)
11674 Support SQL92 syntax for IS TRUE/IS FALSE/IS NOT TRUE/IS NOT FALSE(Thomas)
11675 Allow shorter strings for boolean literals (e.g. "t", "tr", "tru")(Thomas)
11676 Allow SQL92 delimited identifiers(Thomas)
11677 Implement SQL92 binary and hexadecimal string decoding (b'10' and x'1F')(Thomas)
11678 Support SQL92 syntax for type coercion of literal strings
11679 (e.g. "DATETIME 'now'")(Thomas)
11680 Add conversions for int2, int4, and OID types to and from text(Thomas)
11681 Use shared lock when building indexes(Vadim)
11682 Free memory allocated for an user query inside transaction block after
11683 this query is done, was turned off in <= 6.2.1(Vadim)
11684 New SQL statement CREATE PROCEDURAL LANGUAGE(Jan)
11685 New <productname>PostgreSQL</productname> Procedural Language (PL) backend interface(Jan)
11686 Rename pg_dump -H option to -h(Bruce)
11687 Add Java support for passwords, European dates(Peter)
11688 Use indexes for LIKE and ~, !~ operations(Bruce)
11689 Add hash functions for datetime and timespan(Thomas)
11690 Time Travel removed(Vadim, Bruce)
11691 Add paging for \d and \z, and fix \i(Bruce)
11692 Add Unix domain socket support to backend and to frontend library(Goran)
11693 Implement CREATE DATABASE/WITH LOCATION and initlocation utility(Thomas)
11694 Allow more SQL92 and/or <productname>PostgreSQL</productname> reserved words as column identifiers(Thomas)
11695 Augment support for SQL92 SET TIME ZONE...(Thomas)
11696 SET/SHOW/RESET TIME ZONE uses TZ backend environment variable(Thomas)
11697 Implement SET keyword = DEFAULT and SET TIME ZONE DEFAULT(Thomas)
11698 Enable SET TIME ZONE using TZ environment variable(Thomas)
11699 Add PGDATESTYLE environment variable to frontend and backend initialization(Thomas)
11700 Add PGTZ, PGCOSTHEAP, PGCOSTINDEX, PGRPLANS, PGGEQO
11701 frontend library initialization environment variables(Thomas)
11702 Regression tests time zone automatically set with "setenv PGTZ PST8PDT"(Thomas)
11703 Add pg_description table for info on tables, columns, operators, types, and
11705 Increase 16 char limit on system table/index names to 32 characters(Bruce)
11706 Rename system indexes(Bruce)
11707 Add 'GERMAN' option to SET DATESTYLE(Thomas)
11708 Define an "ISO-style" timespan output format with "hh:mm:ss" fields(Thomas)
11709 Allow fractional values for delta times (e.g. '2.5 days')(Thomas)
11710 Validate numeric input more carefully for delta times(Thomas)
11711 Implement day of year as possible input to date_part()(Thomas)
11712 Define timespan_finite() and text_timespan() functions(Thomas)
11713 Remove archive stuff(Bruce)
11714 Allow for a pg_password authentication database that is separate from
11715 the system password file(Todd)
11716 Dump ACLs, GRANT, REVOKE privileges(Matt)
11717 Define text, varchar, and bpchar string length functions(Thomas)
11718 Fix Query handling for inheritance, and cost computations(Bruce)
11719 Implement CREATE TABLE/AS SELECT (alternative to SELECT/INTO)(Thomas)
11720 Allow NOT, IS NULL, IS NOT NULL in constraints(Thomas)
11721 Implement UNIONs for SELECT(Bruce)
11722 Add UNION, GROUP, DISTINCT to INSERT(Bruce)
11723 varchar() stores only necessary bytes on disk(Bruce)
11724 Fix for BLOBs(Peter)
11725 Mega-Patch for JDBC...see README_6.3 for list of changes(Peter)
11726 Remove unused "option" from PQconnectdb()
11727 New LOCK command and lock manual page describing deadlocks(Bruce)
11728 Add new psql \da, \dd, \df, \do, \dS, and \dT commands(Bruce)
11729 Enhance psql \z to show sequences(Bruce)
11730 Show NOT NULL and DEFAULT in psql \d table(Bruce)
11731 New psql .psqlrc file start-up(Andrew)
11732 Modify sample start-up script in contrib/linux to show syslog(Thomas)
11733 New types for IP and MAC addresses in contrib/ip_and_mac(TomH)
11734 Unix system time conversions with date/time types in contrib/unixdate(Thomas)
11735 Update of contrib stuff(Massimo)
11736 Add Unix socket support to DBD::Pg(Goran)
11737 New python interface (PyGreSQL 2.0)(D'Arcy)
11738 New frontend/backend protocol has a version number, network byte order(Phil)
11739 Security features in pg_hba.conf enhanced and documented, many cleanups(Phil)
11740 CHAR() now faster access than VARCHAR() or TEXT
11741 ecpg embedded SQL preprocessor
11742 Reduce system column overhead(Vadmin)
11743 Remove pg_time table(Vadim)
11744 Add pg_type attribute to identify types that need length (bpchar, varchar)
11745 Add report of offending line when COPY command fails
11746 Allow VIEW privileges to be set separately from the underlying tables.
11747 For security, use GRANT/REVOKE on views as appropriate(Jan)
11748 Tables now have no default GRANT SELECT TO PUBLIC. You must
11749 explicitly grant such privileges.
11750 Clean up tutorial examples(Darren)
11752 Source Tree Changes
11753 -------------------
11754 Add new html development tools, and flow chart in /tools/backend
11755 Fix for SCO compiles
11756 Stratus computer port Robert Gillies
11757 Added support for shlib for BSD44_derived & i386_solaris
11758 Make configure more automated(Brook)
11759 Add script to check regression test results
11760 Break parser functions into smaller files, group together(Bruce)
11761 Rename heap_create to heap_create_and_catalog, rename heap_creatr
11762 to heap_create()(Bruce)
11763 Sparc/Linux patch for locking(TomS)
11764 Remove PORTNAME and reorganize port-specific stuff(Marc)
11765 Add optimizer README file(Bruce)
11766 Remove some recursion in optimizer and clean up some code there(Bruce)
11767 Fix for NetBSD locking(Henry)
11768 Fix for libptcl make(Tatsuo)
11770 Change IS TRUE, IS FALSE, ... to expressions using "=" rather than
11771 function calls to istrue() or isfalse() to allow optimization(Thomas)
11772 Various fixes NetBSD/Sparc related(TomH)
11773 Alpha linux locking(Travis,Ryan)
11774 Change elog(WARN) to elog(ERROR)(Bruce)
11775 FAQ for FreeBSD(Marc)
11776 Bring in the PostODBC source tree as part of our standard distribution(Marc)
11777 A minor patch for HP/UX 10 vs 9(Stan)
11778 New pg_attribute.atttypmod for type-specific info like varchar length(Bruce)
11779 UnixWare patches(Billy)
11780 New i386 'lock' for spinlock asm(Billy)
11781 Support for multiplexed backends is removed
11782 Start an OpenBSD port
11784 Start a Cygnus port
11785 Add string functions to regression suite(Thomas)
11786 Expand a few function names formerly truncated to 16 characters(Thomas)
11787 Remove un-needed malloc() calls and replace with palloc()(Bruce)
11793 <sect1 id="release-6-2-1">
11794 <title>Release 6.2.1</title>
11797 <title>Release date</title>
11798 <simpara>1997-10-17</simpara>
11802 6.2.1 is a bug-fix and usability release on 6.2.
11810 Allow strings to span lines, per <acronym>SQL92</acronym>.
11816 Include example trigger function for inserting user names on table updates.
11823 This is a minor bug-fix release on 6.2.
11824 For upgrades from pre-6.2 systems, a full dump/reload is required.
11825 Refer to the 6.2 release notes for instructions.
11829 <title>Migration from version 6.2 to version 6.2.1</title>
11832 This is a minor bug-fix release. A dump/reload is not required from version 6.2,
11833 but is required from any release prior to 6.2.
11836 In upgrading from version 6.2, if you choose to dump/reload you will find that
11837 avg(money) is now calculated correctly. All other bug fixes take effect
11838 upon updating the executables.
11841 Another way to avoid dump/reload is to use the following SQL command
11842 from <command>psql</command> to update the existing system table:
11845 update pg_aggregate set aggfinalfn = 'cash_div_flt8'
11846 where aggname = 'avg' and aggbasetype = 790;
11850 This will need to be done to every existing database, including template1.
11855 <title>Changes</title>
11859 Allow TIME and TYPE column names(Thomas)
11860 Allow larger range of true/false as boolean values(Thomas)
11861 Support output of "now" and "current"(Thomas)
11862 Handle DEFAULT with INSERT of NULL properly(Vadim)
11863 Fix for relation reference counts problem in buffer manager(Vadim)
11864 Allow strings to span lines, like ANSI(Thomas)
11865 Fix for backward cursor with ORDER BY(Vadim)
11866 Fix avg(cash) computation(Thomas)
11867 Fix for specifying a column twice in ORDER/GROUP BY(Vadim)
11868 Documented new libpq function to return affected rows, PQcmdTuples(Bruce)
11869 Trigger function for inserting user names for INSERT/UPDATE(Brook Milligan)
11875 <sect1 id="release-6-2">
11876 <title>Release 6.2</title>
11879 <title>Release date</title>
11880 <simpara>1997-10-02</simpara>
11884 A dump/restore is required for those wishing to migrate data from
11885 previous releases of <productname>PostgreSQL</productname>.
11889 <title>Migration from version 6.1 to version 6.2</title>
11892 This migration requires a complete dump of the 6.1 database and a
11893 restore of the database in 6.2.
11896 Note that the <command>pg_dump</command> and <command>pg_dumpall</command> utility from 6.2 should be used
11897 to dump the 6.1 database.
11902 <title>Migration from version 1.<replaceable>x</> to version 6.2</title>
11905 Those migrating from earlier 1.* releases should first upgrade to 1.09
11906 because the COPY output format was improved from the 1.02 release.
11911 <title>Changes</title>
11917 Fix problems with pg_dump for inheritance, sequences, archive tables(Bruce)
11918 Fix compile errors on overflow due to shifts, unsigned, and bad prototypes
11919 from Solaris(Diab Jerius)
11920 Fix bugs in geometric line arithmetic (bad intersection calculations)(Thomas)
11921 Check for geometric intersections at endpoints to avoid rounding ugliness(Thomas)
11922 Catch non-functional delete attempts(Vadim)
11923 Change time function names to be more consistent(Michael Reifenberg)
11924 Check for zero divides(Michael Reifenberg)
11925 Fix very old bug which made rows changed/inserted by a command
11926 visible to the command itself (so we had multiple update of
11927 updated rows, etc.)(Vadim)
11928 Fix for SELECT null, 'fail' FROM pg_am (Patrick)
11929 SELECT NULL as EMPTY_FIELD now allowed(Patrick)
11930 Remove un-needed signal stuff from contrib/pginterface
11931 Fix OR (where x != 1 or x isnull didn't return rows with x NULL) (Vadim)
11932 Fix time_cmp function (Vadim)
11933 Fix handling of functions with non-attribute first argument in
11934 WHERE clauses (Vadim)
11935 Fix GROUP BY when order of entries is different from order
11936 in target list (Vadim)
11937 Fix pg_dump for aggregates without sfunc1 (Vadim)
11941 Default genetic optimizer GEQO parameter is now 8(Bruce)
11942 Allow use parameters in target list having aggregates in functions(Vadim)
11943 Added JDBC driver as an interface(Adrian & Peter)
11944 pg_password utility
11945 Return number of rows inserted/affected by INSERT/UPDATE/DELETE etc.(Vadim)
11946 Triggers implemented with CREATE TRIGGER (SQL3)(Vadim)
11947 SPI (Server Programming Interface) allows execution of queries inside
11948 C-functions (Vadim)
11949 NOT NULL implemented (SQL92)(Robson Paniago de Miranda)
11950 Include reserved words for string handling, outer joins, and unions(Thomas)
11951 Implement extended comments ("/* ... */") using exclusive states(Thomas)
11952 Add "//" single-line comments(Bruce)
11953 Remove some restrictions on characters in operator names(Thomas)
11954 DEFAULT and CONSTRAINT for tables implemented (SQL92)(Vadim & Thomas)
11955 Add text concatenation operator and function (SQL92)(Thomas)
11956 Support WITH TIME ZONE syntax (SQL92)(Thomas)
11957 Support INTERVAL unit TO unit syntax (SQL92)(Thomas)
11958 Define types DOUBLE PRECISION, INTERVAL, CHARACTER,
11959 and CHARACTER VARYING (SQL92)(Thomas)
11960 Define type FLOAT(p) and rudimentary DECIMAL(p,s), NUMERIC(p,s) (SQL92)(Thomas)
11961 Define EXTRACT(), POSITION(), SUBSTRING(), and TRIM() (SQL92)(Thomas)
11962 Define CURRENT_DATE, CURRENT_TIME, CURRENT_TIMESTAMP (SQL92)(Thomas)
11963 Add syntax and warnings for UNION, HAVING, INNER and OUTER JOIN (SQL92)(Thomas)
11964 Add more reserved words, mostly for SQL92 compliance(Thomas)
11965 Allow hh:mm:ss time entry for timespan/reltime types(Thomas)
11966 Add center() routines for lseg, path, polygon(Thomas)
11967 Add distance() routines for circle-polygon, polygon-polygon(Thomas)
11968 Check explicitly for points and polygons contained within polygons
11969 using an axis-crossing algorithm(Thomas)
11970 Add routine to convert circle-box(Thomas)
11971 Merge conflicting operators for different geometric data types(Thomas)
11972 Replace distance operator "<===>" with "<->"(Thomas)
11973 Replace "above" operator "!^" with ">^" and "below" operator "!|" with "<^"(Thomas)
11974 Add routines for text trimming on both ends, substring, and string position(Thomas)
11975 Added conversion routines circle(box) and poly(circle)(Thomas)
11976 Allow internal sorts to be stored in memory rather than in files(Bruce & Vadim)
11977 Allow functions and operators on internally-identical types to succeed(Bruce)
11978 Speed up backend start-up after profiling analysis(Bruce)
11979 Inline frequently called functions for performance(Bruce)
11980 Reduce open() calls(Bruce)
11981 psql: Add PAGER for \h and \?,\C fix
11982 Fix for psql pager when no tty(Bruce)
11983 New entab utility(Bruce)
11984 General trigger functions for referential integrity (Vadim)
11985 General trigger functions for time travel (Vadim)
11986 General trigger functions for AUTOINCREMENT/IDENTITY feature (Vadim)
11987 MOVE implementation (Vadim)
11989 Source Tree Changes
11990 -------------------
11991 HP-UX 10 patches (Vladimir Turin)
11992 Added SCO support, (Daniel Harris)
11993 MkLinux patches (Tatsuo Ishii)
11994 Change geometric box terminology from "length" to "width"(Thomas)
11995 Deprecate temporary unstored slope fields in geometric code(Thomas)
11996 Remove restart instructions from INSTALL(Bruce)
11997 Look in /usr/ucb first for install(Bruce)
11998 Fix c++ copy example code(Thomas)
11999 Add -o to psql manual page(Bruce)
12000 Prevent relname unallocated string length from being copied into database(Bruce)
12001 Cleanup for NAMEDATALEN use(Bruce)
12002 Fix pg_proc names over 15 chars in output(Bruce)
12003 Add strNcpy() function(Bruce)
12004 remove some (void) casts that are unnecessary(Bruce)
12005 new interfaces directory(Marc)
12006 Replace fopen() calls with calls to fd.c functions(Bruce)
12007 Make functions static where possible(Bruce)
12008 enclose unused functions in #ifdef NOT_USED(Bruce)
12009 Remove call to difftime() in timestamp support to fix SunOS(Bruce & Thomas)
12010 Changes for Digital Unix
12011 Portability fix for pg_dumpall(Bruce)
12012 Rename pg_attribute.attnvals to attdispersion(Bruce)
12013 "intro/unix" manual page now "pgintro"(Bruce)
12014 "built-in" manual page now "pgbuiltin"(Bruce)
12015 "drop" manual page now "drop_table"(Bruce)
12016 Add "create_trigger", "drop_trigger" manual pages(Thomas)
12017 Add constraints regression test(Vadim & Thomas)
12018 Add comments syntax regression test(Thomas)
12019 Add PGINDENT and support program(Bruce)
12020 Massive commit to run PGINDENT on all *.c and *.h files(Bruce)
12021 Files moved to /src/tools directory(Bruce)
12022 SPI and Trigger programming guides (Vadim & D'Arcy)
12028 <sect1 id="release-6-1-1">
12029 <title>Release 6.1.1</title>
12032 <title>Release date</title>
12033 <simpara>1997-07-22</simpara>
12037 <title>Migration from version 6.1 to version 6.1.1</title>
12040 This is a minor bug-fix release. A dump/reload is not required from version 6.1,
12041 but is required from any release prior to 6.1.
12042 Refer to the release notes for 6.1 for more details.
12047 <title>Changes</title>
12051 fix for SET with options (Thomas)
12052 allow pg_dump/pg_dumpall to preserve ownership of all tables/objects(Bruce)
12053 new psql \connect option allows changing usernames without changing databases
12054 fix for initdb --debug option(Yoshihiko Ichikawa))
12055 lextest cleanup(Bruce)
12057 fix date/time month boundary arithmetic(Thomas)
12058 fix timezone daylight handling for some ports(Thomas, Bruce, Tatsuo)
12059 timestamp overhauled to use standard functions(Thomas)
12060 other code cleanup in date/time routines(Thomas)
12061 psql's \d now case-insensitive(Bruce)
12062 psql's backslash commands can now have trailing semicolon(Bruce)
12063 fix memory leak in psql when using \g(Bruce)
12064 major fix for endian handling of communication to server(Thomas, Tatsuo)
12065 Fix for Solaris assembler and include files(Yoshihiko Ichikawa)
12066 allow underscores in usernames(Bruce)
12067 pg_dumpall now returns proper status, portability fix(Bruce)
12073 <sect1 id="release-6-1">
12074 <title>Release 6.1</title>
12077 <title>Release date</title>
12078 <simpara>1997-06-08</simpara>
12082 The regression tests have been adapted and extensively modified for the
12083 6.1 release of <productname>PostgreSQL</productname>.
12087 Three new data types (<type>datetime</type>, <type>timespan</type>, and <type>circle</type>) have been added to
12088 the native set of <productname>PostgreSQL</productname> types. Points, boxes, paths, and polygons
12089 have had their output formats made consistent across the data types.
12090 The polygon output in misc.out has only been spot-checked for correctness
12091 relative to the original regression output.
12095 <productname>PostgreSQL</productname> 6.1 introduces a new, alternate
12096 optimizer which uses <firstterm>genetic</firstterm>
12097 algorithms. These algorithms introduce a random behavior in the ordering
12098 of query results when the query contains multiple qualifiers or multiple
12099 tables (giving the optimizer a choice on order of evaluation). Several
12100 regression tests have been modified to explicitly order the results, and
12101 hence are insensitive to optimizer choices. A few regression tests are
12102 for data types which are inherently unordered (e.g. points and time
12103 intervals) and tests involving those types are explicitly bracketed with
12104 <command>set geqo to 'off'</command> and <command>reset geqo</command>.
12108 The interpretation of array specifiers (the curly braces around atomic
12109 values) appears to have changed sometime after the original regression
12110 tests were generated. The current <filename>./expected/*.out</filename> files reflect this
12111 new interpretation, which may not be correct!
12115 The float8 regression test fails on at least some platforms. This is due
12116 to differences in implementations of <function>pow()</function> and <function>exp()</function> and the signaling
12117 mechanisms used for overflow and underflow conditions.
12121 The <quote>random</> results in the random test should cause the
12122 <quote>random</quote> test to be <quote>failed</quote>, since the
12123 regression tests are evaluated using a simple diff. However,
12124 <quote>random</> does not seem to produce random results on my test
12125 machine (Linux/<application>gcc</>/i686).
12129 <title>Migration to version 6.1</title>
12132 This migration requires a complete dump of the 6.0 database and a
12133 restore of the database in 6.1.
12136 Those migrating from earlier 1.* releases should first upgrade to 1.09
12137 because the COPY output format was improved from the 1.02 release.
12142 <title>Changes</title>
12148 packet length checking in library routines
12149 lock manager priority patch
12150 check for under/over flow of float8(Bruce)
12151 multitable join fix(Vadim)
12152 SIGPIPE crash fix(Darren)
12153 large object fixes(Sven)
12154 allow btree indexes to handle NULLs(Vadim)
12155 timezone fixes(D'Arcy)
12156 select SUM(x) can return NULL on no rows(Thomas)
12157 internal optimizer, executor bug fixes(Vadim)
12158 fix problem where inner loop in < or <= has no rows(Vadim)
12159 prevent re-commuting join index clauses(Vadim)
12160 fix join clauses for multiple tables(Vadim)
12161 fix hash, hashjoin for arrays(Vadim)
12162 fix btree for abstime type(Vadim)
12163 large object fixes(Raymond)
12164 fix buffer leak in hash indexes (Vadim)
12165 fix rtree for use in inner scan (Vadim)
12166 fix gist for use in inner scan, cleanups (Vadim, Andrea)
12167 avoid unnecessary local buffers allocation (Vadim, Massimo)
12168 fix local buffers leak in transaction aborts (Vadim)
12169 fix file manager memmory leaks, cleanups (Vadim, Massimo)
12170 fix storage manager memmory leaks (Vadim)
12171 fix btree duplicates handling (Vadim)
12172 fix deleted rows reincarnation caused by vacuum (Vadim)
12173 fix SELECT varchar()/char() INTO TABLE made zero-length fields(Bruce)
12174 many psql, pg_dump, and libpq memory leaks fixed using Purify (Igor)
12178 attribute optimization statistics(Bruce)
12179 much faster new btree bulk load code(Paul)
12180 BTREE UNIQUE added to bulk load code(Vadim)
12181 new lock debug code(Massimo)
12182 massive changes to libpg++(Leo)
12183 new GEQO optimizer speeds table multitable optimization(Martin)
12184 new WARN message for non-unique insert into unique key(Marc)
12185 update x=-3, no spaces, now valid(Bruce)
12186 remove case-sensitive identifier handling(Bruce,Thomas,Dan)
12187 debug backend now pretty-prints tree(Darren)
12188 new Oracle character functions(Edmund)
12189 new plaintext password functions(Dan)
12190 no such class or insufficient privilege changed to distinct messages(Dan)
12191 new ANSI timestamp function(Dan)
12192 new ANSI Time and Date types (Thomas)
12193 move large chunks of data in backend(Martin)
12194 multicolumn btree indexes(Vadim)
12195 new SET var TO value command(Martin)
12196 update transaction status on reads(Dan)
12197 new locale settings for character types(Oleg)
12198 new SEQUENCE serial number generator(Vadim)
12199 GROUP BY function now possible(Vadim)
12200 re-organize regression test(Thomas,Marc)
12201 new optimizer operation weights(Vadim)
12202 new psql \z grant/permit option(Marc)
12203 new MONEY data type(D'Arcy,Thomas)
12204 tcp socket communication speed improved(Vadim)
12205 new VACUUM option for attribute statistics, and for certain columns (Vadim)
12206 many geometric type improvements(Thomas,Keith)
12207 additional regression tests(Thomas)
12208 new datestyle variable(Thomas,Vadim,Martin)
12209 more comparison operators for sorting types(Thomas)
12210 new conversion functions(Thomas)
12211 new more compact btree format(Vadim)
12212 allow pg_dumpall to preserve database ownership(Bruce)
12213 new SET GEQO=# and R_PLANS variable(Vadim)
12214 old (!GEQO) optimizer can use right-sided plans (Vadim)
12215 typechecking improvement in SQL parser(Bruce)
12216 new SET, SHOW, RESET commands(Thomas,Vadim)
12217 new \connect database USER option
12218 new destroydb -i option (Igor)
12219 new \dt and \di psql commands (Darren)
12220 SELECT "\n" now escapes newline (A. Duursma)
12221 new geometry conversion functions from old format (Thomas)
12223 Source tree changes
12224 -------------------
12225 new configuration script(Marc)
12226 readline configuration option added(Marc)
12227 OS-specific configuration options removed(Marc)
12228 new OS-specific template files(Marc)
12229 no more need to edit Makefile.global(Marc)
12230 re-arrange include files(Marc)
12231 nextstep patches (Gregor Hoffleit)
12232 removed Windows-specific code(Bruce)
12233 removed postmaster -e option, now only postgres -e option (Bruce)
12234 merge duplicate library code in front/backends(Martin)
12235 now works with eBones, international Kerberos(Jun)
12236 more shared library support
12237 c++ include file cleanup(Bruce)
12238 warn about buggy flex(Bruce)
12239 DG/UX, Ultrix, IRIX, AIX portability fixes
12245 <sect1 id="release-6-0">
12246 <title>Release 6.0</title>
12249 <title>Release date</title>
12250 <simpara>1997-01-29</simpara>
12254 A dump/restore is required for those wishing to migrate data from
12255 previous releases of <productname>PostgreSQL</productname>.
12259 <title>Migration from version 1.09 to version 6.0</title>
12262 This migration requires a complete dump of the 1.09 database and a
12263 restore of the database in 6.0.
12268 <title>Migration from pre-1.09 to version 6.0</title>
12271 Those migrating from earlier 1.* releases should first upgrade to 1.09
12272 because the COPY output format was improved from the 1.02 release.
12277 <title>Changes</title>
12283 ALTER TABLE bug - running postgress process needs to re-read table definition
12284 Allow vacuum to be run on one table or entire database(Bruce)
12286 Fix array over-runs of memory writes(Kurt)
12287 Fix elusive btree range/non-range bug(Dan)
12288 Fix for hash indexes on some types like time and date
12289 Fix for pg_log size explosion
12290 Fix permissions on lo_export()(Bruce)
12291 Fix unitialized reads of memory(Kurt)
12292 Fixed ALTER TABLE ... char(3) bug(Bruce)
12293 Fixed a few small memory leaks
12294 Fixed EXPLAIN handling of options and changed full_path option name
12295 Fixed output of group acl privileges
12296 Memory leaks (hunt and destroy with tools like Purify(Kurt)
12297 Minor improvements to rules system
12299 New asserts for run-checking
12300 Overhauled parser/analyze code to properly report errors and increase speed
12301 Pg_dump -d now handles NULL's properly(Bruce)
12302 Prevent SELECT NULL from crashing server (Bruce)
12303 Properly report errors when INSERT ... SELECT columns did not match
12304 Properly report errors when insert column names were not correct
12305 psql \g filename now works(Bruce)
12306 psql fixed problem with multiple statements on one line with multiple outputs
12307 Removed duplicate system OIDs
12308 SELECT * INTO TABLE . GROUP/ORDER BY gives unlink error if table exists(Bruce)
12309 Several fixes for queries that crashed the backend
12310 Starting quote in insert string errors(Bruce)
12311 Submitting an empty query now returns empty status, not just " " query(Bruce)
12315 Add EXPLAIN manual page(Bruce)
12316 Add UNIQUE index capability(Dan)
12317 Add hostname/user level access control rather than just hostname and user
12318 Add synonym of != for <>(Bruce)
12319 Allow "select oid,* from table"
12320 Allow BY,ORDER BY to specify columns by number, or by non-alias table.column(Bruce)
12321 Allow COPY from the frontend(Bryan)
12322 Allow GROUP BY to use alias column name(Bruce)
12323 Allow actual compression, not just reuse on the same page(Vadim)
12324 Allow installation-configuration option to auto-add all local users(Bryan)
12325 Allow libpq to distinguish between text value '' and null(Bruce)
12326 Allow non-postgres users with createdb privs to destroydb's
12327 Allow restriction on who can create C functions(Bryan)
12328 Allow restriction on who can do backend COPY(Bryan)
12329 Can shrink tables, pg_time and pg_log(Vadim & Erich)
12330 Change debug level 2 to print queries only, changed debug heading layout(Bruce)
12331 Change default decimal constant representation from float4 to float8(Bruce)
12332 European date format now set when postmaster is started
12333 Execute lowercase function names if not found with exact case
12334 Fixes for aggregate/GROUP processing, allow 'select sum(func(x),sum(x+y) from z'
12335 Gist now included in the distrubution(Marc)
12336 Idend authentication of local users(Bryan)
12337 Implement BETWEEN qualifier(Bruce)
12338 Implement IN qualifier(Bruce)
12339 libpq has PQgetisnull()(Bruce)
12340 libpq++ improvements
12341 New options to initdb(Bryan)
12342 Pg_dump allow dump of OIDs(Bruce)
12343 Pg_dump create indexes after tables are loaded for speed(Bruce)
12344 Pg_dumpall dumps all databases, and the user table
12345 Pginterface additions for NULL values(Bruce)
12346 Prevent postmaster from being run as root
12347 psql \h and \? is now readable(Bruce)
12348 psql allow backslashed, semicolons anywhere on the line(Bruce)
12349 psql changed command prompt for lines in query or in quotes(Bruce)
12350 psql char(3) now displays as (bp)char in \d output(Bruce)
12351 psql return code now more accurate(Bryan?)
12352 psql updated help syntax(Bruce)
12353 Re-visit and fix vacuum(Vadim)
12354 Reduce size of regression diffs, remove timezone name difference(Bruce)
12355 Remove compile-time parameters to enable binary distributions(Bryan)
12356 Reverse meaning of HBA masks(Bryan)
12357 Secure Authentication of local users(Bryan)
12358 Speed up vacuum(Vadim)
12359 Vacuum now had VERBOSE option(Bruce)
12361 Source tree changes
12362 -------------------
12363 All functions now have prototypes that are compared against the calls
12364 Allow asserts to be disabled easly from Makefile.global(Bruce)
12365 Change oid constants used in code to #define names
12366 Decoupled sparc and solaris defines(Kurt)
12367 Gcc -Wall compiles cleanly with warnings only from unfixable constructs
12368 Major include file reorganization/reduction(Marc)
12369 Make now stops on compile failure(Bryan)
12370 Makefile restructuring(Bryan, Marc)
12371 Merge bsdi_2_1 to bsdi(Bruce)
12372 Monitor program removed
12373 Name change from Postgres95 to PostgreSQL
12374 New config.h file(Marc, Bryan)
12375 PG_VERSION now set to 6.0 and used by postmaster
12376 Portability additions, including Ultrix, DG/UX, AIX, and Solaris
12377 Reduced the number of #define's, centeralized #define's
12378 Remove duplicate OIDS in system tables(Dan)
12379 Remove duplicate system catalog info or report mismatches(Dan)
12380 Removed many os-specific #define's
12381 Restructured object file generation/location(Bryan, Marc)
12382 Restructured port-specific file locations(Bryan, Marc)
12383 Unused/uninialized variables corrected
12389 <sect1 id="release-1-09">
12390 <title>Release 1.09</title>
12393 <title>Release date</title>
12394 <simpara>1996-11-04</simpara>
12398 Sorry, we didn't keep track of changes from 1.02 to 1.09. Some of
12399 the changes listed in 6.0 were actually included in the 1.02.1 to 1.09
12404 <sect1 id="release-1-02">
12405 <title>Release 1.02</title>
12408 <title>Release date</title>
12409 <simpara>1996-08-01</simpara>
12413 <title>Migration from version 1.02 to version 1.02.1</title>
12416 Here is a new migration file for 1.02.1. It includes the 'copy' change
12417 and a script to convert old <acronym>ASCII</acronym> files.
12421 The following notes are for the benefit of users who want to migrate
12422 databases from <productname>Postgres95</> 1.01 and 1.02 to <productname>Postgres95</> 1.02.1.
12425 If you are starting afresh with <productname>Postgres95</> 1.02.1 and do not need
12426 to migrate old databases, you do not need to read any further.
12431 In order to upgrade older <productname>Postgres95</> version 1.01 or 1.02 databases to
12432 version 1.02.1, the following steps are required:
12437 Start up a new 1.02.1 postmaster
12442 Add the new built-in functions and operators of 1.02.1 to 1.01 or 1.02
12443 databases. This is done by running the new 1.02.1 server against
12444 your own 1.01 or 1.02 database and applying the queries attached at
12445 the end of the file. This can be done easily through <command>psql</>. If your
12446 1.01 or 1.02 database is named <literal>testdb</literal> and you have cut the commands
12447 from the end of this file and saved them in <filename>addfunc.sql</filename>:
12449 % psql testdb -f addfunc.sql
12452 Those upgrading 1.02 databases will get a warning when executing the
12453 last two statements in the file because they are already present in 1.02. This is
12454 not a cause for concern.
12461 <title>Dump/Reload Procedure</title>
12464 If you are trying to reload a pg_dump or text-mode, <literal>copy tablename to
12465 stdout</literal> generated with a previous version, you will need to run the
12466 attached <command>sed</command> script on the ASCII file before loading it into the
12467 database. The old format used '.' as end-of-data, while '\.' is now the
12468 end-of-data marker. Also, empty strings are now loaded in as '' rather
12469 than NULL. See the copy manual page for full details.
12472 sed 's/^\.$/\\./g' <in_file >out_file
12476 If you are loading an older binary copy or non-<systemitem>stdout</> copy, there is no
12477 end-of-data character, and hence no conversion necessary.
12480 -- following lines added by agc to reflect the case-insensitive
12481 -- regexp searching for varchar (in 1.02), and bpchar (in 1.02.1)
12482 create operator ~* (leftarg = bpchar, rightarg = text, procedure = texticregexeq);
12483 create operator !~* (leftarg = bpchar, rightarg = text, procedure = texticregexne);
12484 create operator ~* (leftarg = varchar, rightarg = text, procedure = texticregexeq);
12485 create operator !~* (leftarg = varchar, rightarg = text, procedure = texticregexne);
12491 <title>Changes</title>
12495 Source code maintenance and development
12496 * worldwide team of volunteers
12497 * the source tree now in CVS at ftp.ki.net
12500 * psql (and underlying libpq library) now has many more options for
12501 formatting output, including HTML
12502 * pg_dump now output the schema and/or the data, with many fixes to
12503 enhance completeness.
12504 * psql used in place of monitor in administration shell scripts.
12505 monitor to be deprecated in next release.
12506 * date/time functions enhanced
12507 * NULL insert/update/comparison fixed/enhanced
12508 * TCL/TK lib and shell fixed to work with both tck7.4/tk4.0 and tcl7.5/tk4.1
12510 Bug Fixes (almost too numerous to mention)
12512 * storage management
12513 * check for NULL pointer before dereferencing
12517 * added SolarisX86 port
12518 * added BSD/OS 2.1 port
12523 Contributors (appologies to any missed)
12524 * Kurt J. Lidl <lidl@va.pubnix.com>
12525 (missed in first run, but no less important)
12526 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
12527 * Jason Wright <jason@shiloh.vnet.net>
12528 * Cees de Groot <C.deGroot@inter.NL.net>
12529 * ernst.molitor@uni-bonn.de
12530 * michael.siebenborn@ae3.Hypo.DE (Michael Siebenborn (6929))
12531 * Brian E. Gallew <geek+@cmu.edu>
12532 * Vadim B. Mikheev <vadim@sable.krasnoyarsk.su>
12533 * Adam Sussman <myddryn@vidya.com>
12534 * Chris Dunlop <chris@onthe.net.au>
12535 * Marc G. Fournier <scrappy@ki.net>
12536 * Dan McGuirk <mcguirk@indirect.com>
12537 * Dr_George_D_Detlefsen <drgeorge@ilt.com>
12538 * Erich Stamberger <eberger@gewi.kfunigraz.ac.at>
12539 * Massimo Dal Zotto <dz@cs.unitn.it>
12540 * Randy Kunkee <kunkee@Starbase.NeoSoft.COM>
12541 * Rick Weldon <rick@wisetech.com>
12542 * Thomas van Reimersdahl <reimersd@dali.techinfo.rwth-aachen.de>
12543 * david bennett <dave@bensoft.com>
12544 * ernst.molitor@uni-bonn.de
12545 * Julian Assange <proff@suburbia.net>
12546 * Bruce Momjian <pgman@candle.pha.pa.us>
12547 * Paul "Shag" Walmsley <ccshag@cclabs.missouri.edu>
12548 * "Alistair G. Crooks" <azcb0@sde.uts.amdahl.com>
12553 <sect1 id="release-1-01">
12554 <title>Release 1.01</title>
12557 <title>Release date</title>
12558 <simpara>1996-02-23</simpara>
12563 <title>Migration from version 1.0 to version 1.01</title>
12566 The following notes are for the benefit of users who want to migrate
12567 databases from <productname>Postgres95</> 1.0 to <productname>Postgres95</> 1.01.
12570 If you are starting afresh with <productname>Postgres95</> 1.01 and do not need
12571 to migrate old databases, you do not need to read any further.
12574 In order to <productname>Postgres95</> version 1.01 with databases created with
12575 <productname>Postgres95</> version 1.0, the following steps are required:
12580 Set the definition of <symbol>NAMEDATALEN</symbol> in <filename>src/Makefile.global</filename> to 16
12581 and <symbol>OIDNAMELEN</symbol> to 20.
12586 Decide whether you want to use Host based authentication.
12591 If you do, you must create a file name <literal>pg_hba</literal> in your top-level data
12592 directory (typically the value of your <envar>$PGDATA</envar>). <filename>src/libpq/pg_hba</filename>
12593 shows an example syntax.
12598 If you do not want host-based authentication, you can comment out
12603 in <filename>src/Makefile.global</filename>
12606 Note that host-based authentication is turned on by default, and if
12607 you do not take steps A or B above, the out-of-the-box 1.01 will
12608 not allow you to connect to 1.0 databases.
12616 Compile and install 1.01, but DO NOT do the <command>initdb</command> step.
12621 Before doing anything else, terminate your 1.0 postmaster, and
12622 backup your existing <envar>$PGDATA</envar> directory.
12627 Set your <envar>PGDATA</envar> environment variable to your 1.0 databases, but set up
12628 path up so that 1.01 binaries are being used.
12633 Modify the file <filename><envar>$PGDATA</envar>/PG_VERSION</filename> from 5.0 to 5.1
12638 Start up a new 1.01 postmaster
12643 Add the new built-in functions and operators of 1.01 to 1.0
12644 databases. This is done by running the new 1.01 server against
12645 your own 1.0 database and applying the queries attached and saving
12646 in the file 1.0_to_1.01.sql. This can be done easily through <command>psql</command>.
12647 If your 1.0 database is name <literal>testdb</literal>:
12650 % psql testdb -f 1.0_to_1.01.sql
12653 and then execute the following commands (cut and paste from here):
12656 -- add builtin functions that are new to 1.01
12658 create function int4eqoid (int4, oid) returns bool as 'foo'
12659 language 'internal';
12660 create function oideqint4 (oid, int4) returns bool as 'foo'
12661 language 'internal';
12662 create function char2icregexeq (char2, text) returns bool as 'foo'
12663 language 'internal';
12664 create function char2icregexne (char2, text) returns bool as 'foo'
12665 language 'internal';
12666 create function char4icregexeq (char4, text) returns bool as 'foo'
12667 language 'internal';
12668 create function char4icregexne (char4, text) returns bool as 'foo'
12669 language 'internal';
12670 create function char8icregexeq (char8, text) returns bool as 'foo'
12671 language 'internal';
12672 create function char8icregexne (char8, text) returns bool as 'foo'
12673 language 'internal';
12674 create function char16icregexeq (char16, text) returns bool as 'foo'
12675 language 'internal';
12676 create function char16icregexne (char16, text) returns bool as 'foo'
12677 language 'internal';
12678 create function texticregexeq (text, text) returns bool as 'foo'
12679 language 'internal';
12680 create function texticregexne (text, text) returns bool as 'foo'
12681 language 'internal';
12683 -- add builtin functions that are new to 1.01
12685 create operator = (leftarg = int4, rightarg = oid, procedure = int4eqoid);
12686 create operator = (leftarg = oid, rightarg = int4, procedure = oideqint4);
12687 create operator ~* (leftarg = char2, rightarg = text, procedure = char2icregexeq);
12688 create operator !~* (leftarg = char2, rightarg = text, procedure = char2icregexne);
12689 create operator ~* (leftarg = char4, rightarg = text, procedure = char4icregexeq);
12690 create operator !~* (leftarg = char4, rightarg = text, procedure = char4icregexne);
12691 create operator ~* (leftarg = char8, rightarg = text, procedure = char8icregexeq);
12692 create operator !~* (leftarg = char8, rightarg = text, procedure = char8icregexne);
12693 create operator ~* (leftarg = char16, rightarg = text, procedure = char16icregexeq);
12694 create operator !~* (leftarg = char16, rightarg = text, procedure = char16icregexne);
12695 create operator ~* (leftarg = text, rightarg = text, procedure = texticregexeq);
12696 create operator !~* (leftarg = text, rightarg = text, procedure = texticregexne);
12704 <title>Changes</title>
12709 * 1.01 is backwards compatible with 1.0 database provided the user
12710 follow the steps outlined in the MIGRATION_from_1.0_to_1.01 file.
12711 If those steps are not taken, 1.01 is not compatible with 1.0 database.
12714 * added PQdisplayTuples() to libpq and changed monitor and psql to use it
12715 * added NeXT port (requires SysVIPC implementation)
12716 * added CAST .. AS ... syntax
12717 * added ASC and DESC key words
12718 * added 'internal' as a possible language for CREATE FUNCTION
12719 internal functions are C functions which have been statically linked
12720 into the postgres backend.
12721 * a new type "name" has been added for system identifiers (table names,
12722 attribute names, etc.) This replaces the old char16 type. The
12723 of name is set by the NAMEDATALEN #define in src/Makefile.global
12724 * a readable reference manual that describes the query language.
12725 * added host-based access control. A configuration file ($PGDATA/pg_hba)
12726 is used to hold the configuration data. If host-based access control
12727 is not desired, comment out HBA=1 in src/Makefile.global.
12728 * changed regex handling to be uniform use of Henry Spencer's regex code
12729 regardless of platform. The regex code is included in the distribution
12730 * added functions and operators for case-insensitive regular expressions.
12731 The operators are ~* and !~*.
12732 * pg_dump uses COPY instead of SELECT loop for better performance
12735 * fixed an optimizer bug that was causing core dumps when
12736 functions calls were used in comparisons in the WHERE clause
12737 * changed all uses of getuid to geteuid so that effective uids are used
12738 * psql now returns non-zero status on errors when using -c
12739 * applied public patches 1-14
12745 <sect1 id="release-1-0">
12746 <title>Release 1.0</title>
12749 <title>Release date</title>
12750 <simpara>1995-09-05</simpara>
12754 <title>Changes</title>
12759 * The copyright of <productname>Postgres</productname> 1.0 has been loosened to be freely modifiable
12760 and modifiable for any purpose. Please read the COPYRIGHT file.
12761 Thanks to Professor Michael Stonebraker for making this possible.
12764 * date formats have to be MM-DD-YYYY (or DD-MM-YYYY if you're using
12765 EUROPEAN STYLE). This follows SQL-92 specs.
12766 * "delimiters" is now a key word
12769 * sql LIKE syntax has been added
12770 * copy command now takes an optional USING DELIMITER specification.
12771 delimiters can be any single-character string.
12772 * IRIX 5.3 port has been added.
12773 Thanks to Paul Walmsley and others.
12774 * updated pg_dump to work with new libpq
12775 * \d has been added psql
12776 Thanks to Keith Parks
12777 * regexp performance for architectures that use POSIX regex has been
12778 improved due to caching of precompiled patterns.
12779 Thanks to Alistair Crooks
12780 * a new version of libpq++
12781 Thanks to William Wanders
12784 * arbitrary userids can be specified in the createuser script
12785 * \c to connect to other databases in psql now works.
12786 * bad pg_proc entry for float4inc() is fixed
12787 * users with usecreatedb field set can now create databases without
12788 having to be usesuper
12789 * remove access control entries when the entry no longer has any
12791 * fixed non-portable datetimes implementation
12792 * added kerberos flags to the src/backend/Makefile
12793 * libpq now works with kerberos
12794 * typographic errors in the user manual have been corrected.
12795 * btrees with multiple index never worked, now we tell you they don't
12796 work when you try to use them
12802 <sect1 id="release-0-03">
12803 <title><productname>Postgres95</productname> Release 0.03</title>
12806 <title>Release date</title>
12807 <simpara>1995-07-21</simpara>
12811 <title>Changes</title>
12814 Incompatible changes:
12815 * BETA-0.3 IS INCOMPATIBLE WITH DATABASES CREATED WITH PREVIOUS VERSIONS
12816 (due to system catalog changes and indexing structure changes).
12817 * double-quote (") is deprecated as a quoting character for string literals;
12818 you need to convert them to single quotes ('). <!-- " -->
12819 * name of aggregates (eg. int4sum) are renamed in accordance with the
12820 SQL standard (eg. sum).
12821 * CHANGE ACL syntax is replaced by GRANT/REVOKE syntax.
12822 * float literals (eg. 3.14) are now of type float4 (instead of float8 in
12823 previous releases); you might have to do typecasting if you depend on it
12824 being of type float8. If you neglect to do the typecasting and you assign
12825 a float literal to a field of type float8, you may get incorrect values
12827 * LIBPQ has been totally revamped so that frontend applications
12828 can connect to multiple backends
12829 * the usesysid field in pg_user has been changed from int2 to int4 to
12830 allow wider range of Unix user ids.
12831 * the netbsd/freebsd/bsd o/s ports have been consolidated into a
12832 single BSD44_derived port. (thanks to Alistair Crooks)
12834 SQL standard-compliance (the following details changes that makes postgres95
12835 more compliant to the SQL-92 standard):
12836 * the following SQL types are now built-in: smallint, int(eger), float, real,
12837 char(N), varchar(N), date and time.
12839 The following are aliases to existing postgres types:
12840 smallint -> int2
12841 integer, int -> int4
12842 float, real -> float4
12843 char(N) and varchar(N) are implemented as truncated text types. In
12844 addition, char(N) does blank-padding.
12845 * single-quote (') is used for quoting string literals; '' (in addition to
12846 \') is supported as means of inserting a single quote in a string
12847 * SQL standard aggregate names (MAX, MIN, AVG, SUM, COUNT) are used
12848 (Also, aggregates can now be overloaded, i.e. you can define your
12849 own MAX aggregate to take in a user-defined type.)
12850 * CHANGE ACL removed. GRANT/REVOKE syntax added.
12851 - Privileges can be given to a group using the "GROUP" key word.
12853 GRANT SELECT ON foobar TO GROUP my_group;
12854 The key word 'PUBLIC' is also supported to mean all users.
12856 Privileges can only be granted or revoked to one user or group
12859 "WITH GRANT OPTION" is not supported. Only class owners can change
12861 - The default access control is to to grant users readonly access.
12862 You must explicitly grant insert/update access to users. To change
12863 this, modify the line in
12864 src/backend/utils/acl.h
12865 that defines ACL_WORLD_DEFAULT
12868 * the bug where aggregates of empty tables were not run has been fixed. Now,
12869 aggregates run on empty tables will return the initial conditions of the
12870 aggregates. Thus, COUNT of an empty table will now properly return 0.
12871 MAX/MIN of an empty table will return a row of value NULL.
12872 * allow the use of \; inside the monitor
12873 * the LISTEN/NOTIFY asynchronous notification mechanism now work
12874 * NOTIFY in rule action bodies now work
12875 * hash indexes work, and access methods in general should perform better.
12876 creation of large btree indexes should be much faster. (thanks to Paul
12879 Other changes and enhancements:
12880 * addition of an EXPLAIN statement used for explaining the query execution
12881 plan (eg. "EXPLAIN SELECT * FROM EMP" prints out the execution plan for
12883 * WARN and NOTICE messages no longer have timestamps on them. To turn on
12884 timestamps of error messages, uncomment the line in
12885 src/backend/utils/elog.h:
12886 /* define ELOG_TIMESTAMPS */
12887 * On an access control violation, the message
12888 "Either no such class or insufficient privilege"
12889 will be given. This is the same message that is returned when
12890 a class is not found. This dissuades non-privileged users from
12891 guessing the existence of privileged classes.
12892 * some additional system catalog changes have been made that are not
12893 visible to the user.
12896 * The -oid option has been added to the "pg_result" tcl command.
12897 pg_result -oid returns oid of the last row inserted. If the
12898 last command was not an INSERT, then pg_result -oid returns "".
12899 * the large object interface is available as pg_lo* tcl commands:
12900 pg_lo_open, pg_lo_close, pg_lo_creat, etc.
12902 Portability enhancements and New Ports:
12903 * flex/lex problems have been cleared up. Now, you should be able to use
12904 flex instead of lex on any platforms. We no longer make assumptions of
12905 what lexer you use based on the platform you use.
12906 * The Linux-ELF port is now supported. Various configuration have been
12907 tested: The following configuration is known to work:
12908 kernel 1.2.10, gcc 2.6.3, libc 4.7.2, flex 2.5.2, bison 1.24
12909 with everything in ELF format,
12912 * ipcclean added to the distribution
12913 ipcclean usually does not need to be run, but if your backend crashes
12914 and leaves shared memory segments hanging around, ipcclean will
12915 clean them up for you.
12918 * the user manual has been revised and libpq documentation added.
12924 <sect1 id="release-0-02">
12925 <title><productname>Postgres95</productname> Release 0.02</title>
12928 <title>Release date</title>
12929 <simpara>1995-05-25</simpara>
12933 <title>Changes</title>
12937 Incompatible changes:
12938 * The SQL statement for creating a database is 'CREATE DATABASE' instead
12939 of 'CREATEDB'. Similarly, dropping a database is 'DROP DATABASE' instead
12940 of 'DESTROYDB'. However, the names of the executables 'createdb' and
12941 'destroydb' remain the same.
12944 * pgperl - a Perl (4.036) interface to Postgres95
12945 * pg_dump - a utility for dumping out a postgres database into a
12946 script file containing query commands. The script files are in a ASCII
12947 format and can be used to reconstruct the database, even on other
12948 machines and other architectures. (Also good for converting
12949 a Postgres 4.2 database to Postgres95 database.)
12951 The following ports have been incorporated into postgres95-beta-0.02:
12952 * the NetBSD port by Alistair Crooks
12953 * the AIX port by Mike Tung
12954 * the Windows NT port by Jon Forrest (more stuff but not done yet)
12955 * the Linux ELF port by Brian Gallew
12957 The following bugs have been fixed in postgres95-beta-0.02:
12958 * new lines not escaped in COPY OUT and problem with COPY OUT when first
12960 * cannot type return to use the default user id in createuser
12961 * SELECT DISTINCT on big tables crashes
12962 * Linux installation problems
12963 * monitor doesn't allow use of 'localhost' as PGHOST
12964 * psql core dumps when doing \c or \l
12965 * the "pgtclsh" target missing from src/bin/pgtclsh/Makefile
12966 * libpgtcl has a hard-wired default port number
12967 * SELECT DISTINCT INTO TABLE hangs
12968 * CREATE TYPE doesn't accept 'variable' as the internallength
12969 * wrong result using more than 1 aggregate in a SELECT
12975 <sect1 id="release-0-01">
12976 <title><productname>Postgres95</productname> Release 0.01</title>
12979 <title>Release date</title>
12980 <simpara>1995-05-01</simpara>
12989 <sect1 id="timing-results">
12990 <title>Timing Results</title>
12993 These timing results are from running the regression test with the commands
12996 % cd src/test/regress
12998 % time make runtest
13002 Timing under Linux 2.0.27 seems to have a roughly 5% variation from run
13003 to run, presumably due to the scheduling vagaries of multitasking systems.
13007 <title>Version 6.5</title>
13010 As has been the case for previous releases, timing between
13011 releases is not directly comparable since new regression tests
13012 have been added. In general, 6.5 is faster than previous
13017 Timing with <function>fsync()</function> disabled:
13021 02:00 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
13022 04:38 Sparc Ultra 1 143MHz, 64MB, Solaris 2.6
13027 Timing with <function>fsync()</function> enabled:
13031 04:21 Dual Pentium Pro 180, 224MB, UW-SCSI, Linux 2.0.36, gcc 2.7.2.3 -O2 -m486
13034 For the <systemitem class="osname">Linux</systemitem> system above, using <acronym>UW-SCSI</acronym> disks rather than (older) <acronym>IDE</acronym>
13035 disks leads to a 50% improvement in speed on the regression test.
13040 <title>Version 6.4beta</title>
13043 The times for this release are not directly comparable to those for previous releases
13044 since some additional regression tests have been included.
13045 In general, however, 6.4 should be slightly faster than the previous release (thanks, Bruce!).
13050 02:26 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
13056 <title>Version 6.3</title>
13059 The times for this release are not directly comparable to those for previous releases
13060 since some additional regression tests have been included and some obsolete tests involving
13061 time travel have been removed.
13062 In general, however, 6.3 is substantially faster than previous releases (thanks, Bruce!).
13067 02:30 Dual Pentium Pro 180, 96MB, UW-SCSI, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
13068 04:12 Dual Pentium Pro 180, 96MB, EIDE, Linux 2.0.30, gcc 2.7.2.1 -O2 -m486
13074 <title>Version 6.1</title>
13079 06:12 Pentium Pro 180, 32MB, EIDE, Linux 2.0.30, gcc 2.7.2 -O2 -m486
13080 12:06 P-100, 48MB, Linux 2.0.29, gcc
13081 39:58 Sparc IPC 32MB, Solaris 2.5, gcc 2.7.2.1 -O -g
13089 <!-- Keep this comment at the end of the file
13094 sgml-minimize-attributes:nil
13095 sgml-always-quote-attributes:t
13098 sgml-parent-document:nil
13099 sgml-default-dtd-file:"./reference.ced"
13100 sgml-exposed-tags:nil
13101 sgml-local-catalogs:("/usr/lib/sgml/catalog")
13102 sgml-local-ecat-files:nil